text
stringclasses 136
values | repetition
int64 1
10
|
|---|---|
University of Florida researchers are addressing a critical gap in medical genetic research - ensuring it better represents and benefits people of all backgrounds.
Their work, led by Kiley Graim, Ph.D., an assistant professor in the Department of Computer & Information Science & Engineering, focuses on improving human health by addressing "ancestral bias" in genetic data, a problem that arises when most research is based on data from a single ancestral group. This bias limits advancements in precision medicine, Graim said, and leaves large portions of the global population underserved when it comes to disease treatment and prevention.
To solve this, the team developed PhyloFrame, a machine-learning tool that uses artificial intelligence to account for ancestral diversity in genetic data. With funding support from the National Institutes of Health, the goal is to improve how diseases are predicted, diagnosed, and treated for everyone, regardless of their ancestry. A paper describing the PhyloFrame method and how it showed marked improvements in precision medicine outcomes was published Monday in Nature Communications.
Graim's inspiration to focus on ancestral bias in genomic data evolved from a conversation with a doctor who was frustrated by a study's limited relevance to his diverse patient population. This encounter led her to explore how AI could help bridge the gap in genetic research.
"I thought to myself, 'I can fix that problem,'" said Graim, whose research centers around machine learning and precision medicine and who is trained in population genomics. "If our training data doesn't match our real-world data, we have ways to deal with that using machine learning. They're not perfect, but they can do a lot to address the issue."
By leveraging data from population genomics database gnomAD, PhyloFrame integrates massive databases of healthy human genomes with the smaller datasets specific to diseases used to train precision medicine models. The models it creates are better equipped to handle diverse genetic backgrounds. For example, it can predict the differences between subtypes of diseases like breast cancer and suggest the best treatment for each patient, regardless of patient ancestry.
Processing such massive amounts of data is no small feat. The team uses UF's HiPerGator, one of the most powerful supercomputers in the country, to analyze genomic information from millions of people. For each person, that means processing 3 billion base pairs of DNA.
"I didn't think it would work as well as it did," said Graim, noting that her doctoral student, Leslie Smith, contributed significantly to the study. "What started as a small project using a simple model to demonstrate the impact of incorporating population genomics data has evolved into securing funds to develop more sophisticated models and to refine how populations are defined."
What sets PhyloFrame apart is its ability to ensure predictions remain accurate across populations by considering genetic differences linked to ancestry. This is crucial because most current models are built using data that does not fully represent the world's population. Much of the existing data comes from research hospitals and patients who trust the health care system. This means populations in small towns or those who distrust medical systems are often left out, making it harder to develop treatments that work well for everyone.
She also estimated 97% of the sequenced samples are from people of European ancestry, due, largely, to national and state level funding and priorities, but also due to socioeconomic factors that snowball at different levels – insurance impacts whether people get treated, for example, which impacts how likely they are to be sequenced.
Some other countries, notably China and Japan, have recently been trying to close this gap, and so there is more data from these countries than there had been previously but still nothing like the European data. Poorer populations are generally excluded entirely." Kiley Graim, Ph.D., Assistant Professor, Department of Computer & Information Science & Engineering, University of Florida
Thus, diversity in training data is essential, Graim said.
"We want these models to work for any patient, not just the ones in our studies," she said. "Having diverse training data makes models better for Europeans, too. Having the population genomics data helps prevent models from overfitting, which means that they'll work better for everyone, including Europeans."
Graim believes tools like PhyloFrame will eventually be used in the clinical setting, replacing traditional models to develop treatment plans tailored to individuals based on their genetic makeup. The team's next steps include refining PhyloFrame and expanding its applications to more diseases.
"My dream is to help advance precision medicine through this kind of machine learning method, so people can get diagnosed early and are treated with what works specifically for them and with the fewest side effects," she said. "Getting the right treatment to the right person at the right time is what we're striving for."
Graim's project received funding from the UF College of Medicine Office of Research's AI2 Datathon grant award, which is designed to help researchers and clinicians harness AI tools to improve human health. | 5 |
University of Florida researchers are addressing a critical gap in medical genetic research - ensuring it better represents and benefits people of all backgrounds.
Their work, led by Kiley Graim, Ph.D., an assistant professor in the Department of Computer & Information Science & Engineering, focuses on improving human health by addressing "ancestral bias" in genetic data, a problem that arises when most research is based on data from a single ancestral group. This bias limits advancements in precision medicine, Graim said, and leaves large portions of the global population underserved when it comes to disease treatment and prevention.
To solve this, the team developed PhyloFrame, a machine-learning tool that uses artificial intelligence to account for ancestral diversity in genetic data. With funding support from the National Institutes of Health, the goal is to improve how diseases are predicted, diagnosed, and treated for everyone, regardless of their ancestry. A paper describing the PhyloFrame method and how it showed marked improvements in precision medicine outcomes was published Monday in Nature Communications.
Graim's inspiration to focus on ancestral bias in genomic data evolved from a conversation with a doctor who was frustrated by a study's limited relevance to his diverse patient population. This encounter led her to explore how AI could help bridge the gap in genetic research.
"I thought to myself, 'I can fix that problem,'" said Graim, whose research centers around machine learning and precision medicine and who is trained in population genomics. "If our training data doesn't match our real-world data, we have ways to deal with that using machine learning. They're not perfect, but they can do a lot to address the issue."
By leveraging data from population genomics database gnomAD, PhyloFrame integrates massive databases of healthy human genomes with the smaller datasets specific to diseases used to train precision medicine models. The models it creates are better equipped to handle diverse genetic backgrounds. For example, it can predict the differences between subtypes of diseases like breast cancer and suggest the best treatment for each patient, regardless of patient ancestry.
Processing such massive amounts of data is no small feat. The team uses UF's HiPerGator, one of the most powerful supercomputers in the country, to analyze genomic information from millions of people. For each person, that means processing 3 billion base pairs of DNA.
"I didn't think it would work as well as it did," said Graim, noting that her doctoral student, Leslie Smith, contributed significantly to the study. "What started as a small project using a simple model to demonstrate the impact of incorporating population genomics data has evolved into securing funds to develop more sophisticated models and to refine how populations are defined."
What sets PhyloFrame apart is its ability to ensure predictions remain accurate across populations by considering genetic differences linked to ancestry. This is crucial because most current models are built using data that does not fully represent the world's population. Much of the existing data comes from research hospitals and patients who trust the health care system. This means populations in small towns or those who distrust medical systems are often left out, making it harder to develop treatments that work well for everyone.
She also estimated 97% of the sequenced samples are from people of European ancestry, due, largely, to national and state level funding and priorities, but also due to socioeconomic factors that snowball at different levels – insurance impacts whether people get treated, for example, which impacts how likely they are to be sequenced.
Some other countries, notably China and Japan, have recently been trying to close this gap, and so there is more data from these countries than there had been previously but still nothing like the European data. Poorer populations are generally excluded entirely." Kiley Graim, Ph.D., Assistant Professor, Department of Computer & Information Science & Engineering, University of Florida
Thus, diversity in training data is essential, Graim said.
"We want these models to work for any patient, not just the ones in our studies," she said. "Having diverse training data makes models better for Europeans, too. Having the population genomics data helps prevent models from overfitting, which means that they'll work better for everyone, including Europeans."
Graim believes tools like PhyloFrame will eventually be used in the clinical setting, replacing traditional models to develop treatment plans tailored to individuals based on their genetic makeup. The team's next steps include refining PhyloFrame and expanding its applications to more diseases.
"My dream is to help advance precision medicine through this kind of machine learning method, so people can get diagnosed early and are treated with what works specifically for them and with the fewest side effects," she said. "Getting the right treatment to the right person at the right time is what we're striving for."
Graim's project received funding from the UF College of Medicine Office of Research's AI2 Datathon grant award, which is designed to help researchers and clinicians harness AI tools to improve human health. | 5 |
University of Florida researchers are addressing a critical gap in medical genetic research - ensuring it better represents and benefits people of all backgrounds.
Their work, led by Kiley Graim, Ph.D., an assistant professor in the Department of Computer & Information Science & Engineering, focuses on improving human health by addressing "ancestral bias" in genetic data, a problem that arises when most research is based on data from a single ancestral group. This bias limits advancements in precision medicine, Graim said, and leaves large portions of the global population underserved when it comes to disease treatment and prevention.
To solve this, the team developed PhyloFrame, a machine-learning tool that uses artificial intelligence to account for ancestral diversity in genetic data. With funding support from the National Institutes of Health, the goal is to improve how diseases are predicted, diagnosed, and treated for everyone, regardless of their ancestry. A paper describing the PhyloFrame method and how it showed marked improvements in precision medicine outcomes was published Monday in Nature Communications.
Graim's inspiration to focus on ancestral bias in genomic data evolved from a conversation with a doctor who was frustrated by a study's limited relevance to his diverse patient population. This encounter led her to explore how AI could help bridge the gap in genetic research.
"I thought to myself, 'I can fix that problem,'" said Graim, whose research centers around machine learning and precision medicine and who is trained in population genomics. "If our training data doesn't match our real-world data, we have ways to deal with that using machine learning. They're not perfect, but they can do a lot to address the issue."
By leveraging data from population genomics database gnomAD, PhyloFrame integrates massive databases of healthy human genomes with the smaller datasets specific to diseases used to train precision medicine models. The models it creates are better equipped to handle diverse genetic backgrounds. For example, it can predict the differences between subtypes of diseases like breast cancer and suggest the best treatment for each patient, regardless of patient ancestry.
Processing such massive amounts of data is no small feat. The team uses UF's HiPerGator, one of the most powerful supercomputers in the country, to analyze genomic information from millions of people. For each person, that means processing 3 billion base pairs of DNA.
"I didn't think it would work as well as it did," said Graim, noting that her doctoral student, Leslie Smith, contributed significantly to the study. "What started as a small project using a simple model to demonstrate the impact of incorporating population genomics data has evolved into securing funds to develop more sophisticated models and to refine how populations are defined."
What sets PhyloFrame apart is its ability to ensure predictions remain accurate across populations by considering genetic differences linked to ancestry. This is crucial because most current models are built using data that does not fully represent the world's population. Much of the existing data comes from research hospitals and patients who trust the health care system. This means populations in small towns or those who distrust medical systems are often left out, making it harder to develop treatments that work well for everyone.
She also estimated 97% of the sequenced samples are from people of European ancestry, due, largely, to national and state level funding and priorities, but also due to socioeconomic factors that snowball at different levels – insurance impacts whether people get treated, for example, which impacts how likely they are to be sequenced.
Some other countries, notably China and Japan, have recently been trying to close this gap, and so there is more data from these countries than there had been previously but still nothing like the European data. Poorer populations are generally excluded entirely." Kiley Graim, Ph.D., Assistant Professor, Department of Computer & Information Science & Engineering, University of Florida
Thus, diversity in training data is essential, Graim said.
"We want these models to work for any patient, not just the ones in our studies," she said. "Having diverse training data makes models better for Europeans, too. Having the population genomics data helps prevent models from overfitting, which means that they'll work better for everyone, including Europeans."
Graim believes tools like PhyloFrame will eventually be used in the clinical setting, replacing traditional models to develop treatment plans tailored to individuals based on their genetic makeup. The team's next steps include refining PhyloFrame and expanding its applications to more diseases.
"My dream is to help advance precision medicine through this kind of machine learning method, so people can get diagnosed early and are treated with what works specifically for them and with the fewest side effects," she said. "Getting the right treatment to the right person at the right time is what we're striving for."
Graim's project received funding from the UF College of Medicine Office of Research's AI2 Datathon grant award, which is designed to help researchers and clinicians harness AI tools to improve human health. | 5 |
University of Florida researchers are addressing a critical gap in medical genetic research - ensuring it better represents and benefits people of all backgrounds.
Their work, led by Kiley Graim, Ph.D., an assistant professor in the Department of Computer & Information Science & Engineering, focuses on improving human health by addressing "ancestral bias" in genetic data, a problem that arises when most research is based on data from a single ancestral group. This bias limits advancements in precision medicine, Graim said, and leaves large portions of the global population underserved when it comes to disease treatment and prevention.
To solve this, the team developed PhyloFrame, a machine-learning tool that uses artificial intelligence to account for ancestral diversity in genetic data. With funding support from the National Institutes of Health, the goal is to improve how diseases are predicted, diagnosed, and treated for everyone, regardless of their ancestry. A paper describing the PhyloFrame method and how it showed marked improvements in precision medicine outcomes was published Monday in Nature Communications.
Graim's inspiration to focus on ancestral bias in genomic data evolved from a conversation with a doctor who was frustrated by a study's limited relevance to his diverse patient population. This encounter led her to explore how AI could help bridge the gap in genetic research.
"I thought to myself, 'I can fix that problem,'" said Graim, whose research centers around machine learning and precision medicine and who is trained in population genomics. "If our training data doesn't match our real-world data, we have ways to deal with that using machine learning. They're not perfect, but they can do a lot to address the issue."
By leveraging data from population genomics database gnomAD, PhyloFrame integrates massive databases of healthy human genomes with the smaller datasets specific to diseases used to train precision medicine models. The models it creates are better equipped to handle diverse genetic backgrounds. For example, it can predict the differences between subtypes of diseases like breast cancer and suggest the best treatment for each patient, regardless of patient ancestry.
Processing such massive amounts of data is no small feat. The team uses UF's HiPerGator, one of the most powerful supercomputers in the country, to analyze genomic information from millions of people. For each person, that means processing 3 billion base pairs of DNA.
"I didn't think it would work as well as it did," said Graim, noting that her doctoral student, Leslie Smith, contributed significantly to the study. "What started as a small project using a simple model to demonstrate the impact of incorporating population genomics data has evolved into securing funds to develop more sophisticated models and to refine how populations are defined."
What sets PhyloFrame apart is its ability to ensure predictions remain accurate across populations by considering genetic differences linked to ancestry. This is crucial because most current models are built using data that does not fully represent the world's population. Much of the existing data comes from research hospitals and patients who trust the health care system. This means populations in small towns or those who distrust medical systems are often left out, making it harder to develop treatments that work well for everyone.
She also estimated 97% of the sequenced samples are from people of European ancestry, due, largely, to national and state level funding and priorities, but also due to socioeconomic factors that snowball at different levels – insurance impacts whether people get treated, for example, which impacts how likely they are to be sequenced.
Some other countries, notably China and Japan, have recently been trying to close this gap, and so there is more data from these countries than there had been previously but still nothing like the European data. Poorer populations are generally excluded entirely." Kiley Graim, Ph.D., Assistant Professor, Department of Computer & Information Science & Engineering, University of Florida
Thus, diversity in training data is essential, Graim said.
"We want these models to work for any patient, not just the ones in our studies," she said. "Having diverse training data makes models better for Europeans, too. Having the population genomics data helps prevent models from overfitting, which means that they'll work better for everyone, including Europeans."
Graim believes tools like PhyloFrame will eventually be used in the clinical setting, replacing traditional models to develop treatment plans tailored to individuals based on their genetic makeup. The team's next steps include refining PhyloFrame and expanding its applications to more diseases.
"My dream is to help advance precision medicine through this kind of machine learning method, so people can get diagnosed early and are treated with what works specifically for them and with the fewest side effects," she said. "Getting the right treatment to the right person at the right time is what we're striving for."
Graim's project received funding from the UF College of Medicine Office of Research's AI2 Datathon grant award, which is designed to help researchers and clinicians harness AI tools to improve human health. | 5 |
University of Florida researchers are addressing a critical gap in medical genetic research - ensuring it better represents and benefits people of all backgrounds.
Their work, led by Kiley Graim, Ph.D., an assistant professor in the Department of Computer & Information Science & Engineering, focuses on improving human health by addressing "ancestral bias" in genetic data, a problem that arises when most research is based on data from a single ancestral group. This bias limits advancements in precision medicine, Graim said, and leaves large portions of the global population underserved when it comes to disease treatment and prevention.
To solve this, the team developed PhyloFrame, a machine-learning tool that uses artificial intelligence to account for ancestral diversity in genetic data. With funding support from the National Institutes of Health, the goal is to improve how diseases are predicted, diagnosed, and treated for everyone, regardless of their ancestry. A paper describing the PhyloFrame method and how it showed marked improvements in precision medicine outcomes was published Monday in Nature Communications.
Graim's inspiration to focus on ancestral bias in genomic data evolved from a conversation with a doctor who was frustrated by a study's limited relevance to his diverse patient population. This encounter led her to explore how AI could help bridge the gap in genetic research.
"I thought to myself, 'I can fix that problem,'" said Graim, whose research centers around machine learning and precision medicine and who is trained in population genomics. "If our training data doesn't match our real-world data, we have ways to deal with that using machine learning. They're not perfect, but they can do a lot to address the issue."
By leveraging data from population genomics database gnomAD, PhyloFrame integrates massive databases of healthy human genomes with the smaller datasets specific to diseases used to train precision medicine models. The models it creates are better equipped to handle diverse genetic backgrounds. For example, it can predict the differences between subtypes of diseases like breast cancer and suggest the best treatment for each patient, regardless of patient ancestry.
Processing such massive amounts of data is no small feat. The team uses UF's HiPerGator, one of the most powerful supercomputers in the country, to analyze genomic information from millions of people. For each person, that means processing 3 billion base pairs of DNA.
"I didn't think it would work as well as it did," said Graim, noting that her doctoral student, Leslie Smith, contributed significantly to the study. "What started as a small project using a simple model to demonstrate the impact of incorporating population genomics data has evolved into securing funds to develop more sophisticated models and to refine how populations are defined."
What sets PhyloFrame apart is its ability to ensure predictions remain accurate across populations by considering genetic differences linked to ancestry. This is crucial because most current models are built using data that does not fully represent the world's population. Much of the existing data comes from research hospitals and patients who trust the health care system. This means populations in small towns or those who distrust medical systems are often left out, making it harder to develop treatments that work well for everyone.
She also estimated 97% of the sequenced samples are from people of European ancestry, due, largely, to national and state level funding and priorities, but also due to socioeconomic factors that snowball at different levels – insurance impacts whether people get treated, for example, which impacts how likely they are to be sequenced.
Some other countries, notably China and Japan, have recently been trying to close this gap, and so there is more data from these countries than there had been previously but still nothing like the European data. Poorer populations are generally excluded entirely." Kiley Graim, Ph.D., Assistant Professor, Department of Computer & Information Science & Engineering, University of Florida
Thus, diversity in training data is essential, Graim said.
"We want these models to work for any patient, not just the ones in our studies," she said. "Having diverse training data makes models better for Europeans, too. Having the population genomics data helps prevent models from overfitting, which means that they'll work better for everyone, including Europeans."
Graim believes tools like PhyloFrame will eventually be used in the clinical setting, replacing traditional models to develop treatment plans tailored to individuals based on their genetic makeup. The team's next steps include refining PhyloFrame and expanding its applications to more diseases.
"My dream is to help advance precision medicine through this kind of machine learning method, so people can get diagnosed early and are treated with what works specifically for them and with the fewest side effects," she said. "Getting the right treatment to the right person at the right time is what we're striving for."
Graim's project received funding from the UF College of Medicine Office of Research's AI2 Datathon grant award, which is designed to help researchers and clinicians harness AI tools to improve human health. | 5 |
After a review found that government-funded projects face overly complex spending approval processes that bog them down, the route to funding is set to be simplified.
The changes will see projects supported with larger tests if they have the potential to save money or improve public services.
Subscribe today for free
Peter Kyle, the UK’s Technology Secretary, said: “Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in Government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The research funding review comes as part of a wider push by Sir Keir Starmer’s Labour government to make the UK more attractive to tech investments, which includes creating AI Growth Zones and streamlining planning rules for projects like data centres and nuclear reactors .
“This review will help us build technology that will mean businesses can skip the admin and get on with driving growth, digital systems supporting the police are more reliable so they can keep our streets safe, and it will mean we can build new tools to speed up wait times for doctors’ appointments and get the NHS back on its feet are built,” Kyle added.
The government said four new processes will be tested from April, adding to its experimental generative AI chatbot, GOV.UK Chat.
Darren Jones, Chief Secretary to the Treasury, said: “This Government is determined that digital transformation of the state and our public services will deliver better outcomes for people, and ensure every pound of taxpayers money is spent well.”
RELATED STORIES
Government unveils digital inclusion action plan to tackle digital poverty
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms | 5 |
After a review found that government-funded projects face overly complex spending approval processes that bog them down, the route to funding is set to be simplified.
The changes will see projects supported with larger tests if they have the potential to save money or improve public services.
Subscribe today for free
Peter Kyle, the UK’s Technology Secretary, said: “Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in Government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The research funding review comes as part of a wider push by Sir Keir Starmer’s Labour government to make the UK more attractive to tech investments, which includes creating AI Growth Zones and streamlining planning rules for projects like data centres and nuclear reactors .
“This review will help us build technology that will mean businesses can skip the admin and get on with driving growth, digital systems supporting the police are more reliable so they can keep our streets safe, and it will mean we can build new tools to speed up wait times for doctors’ appointments and get the NHS back on its feet are built,” Kyle added.
The government said four new processes will be tested from April, adding to its experimental generative AI chatbot, GOV.UK Chat.
Darren Jones, Chief Secretary to the Treasury, said: “This Government is determined that digital transformation of the state and our public services will deliver better outcomes for people, and ensure every pound of taxpayers money is spent well.”
RELATED STORIES
Government unveils digital inclusion action plan to tackle digital poverty
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms | 5 |
After a review found that government-funded projects face overly complex spending approval processes that bog them down, the route to funding is set to be simplified.
The changes will see projects supported with larger tests if they have the potential to save money or improve public services.
Subscribe today for free
Peter Kyle, the UK’s Technology Secretary, said: “Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in Government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The research funding review comes as part of a wider push by Sir Keir Starmer’s Labour government to make the UK more attractive to tech investments, which includes creating AI Growth Zones and streamlining planning rules for projects like data centres and nuclear reactors .
“This review will help us build technology that will mean businesses can skip the admin and get on with driving growth, digital systems supporting the police are more reliable so they can keep our streets safe, and it will mean we can build new tools to speed up wait times for doctors’ appointments and get the NHS back on its feet are built,” Kyle added.
The government said four new processes will be tested from April, adding to its experimental generative AI chatbot, GOV.UK Chat.
Darren Jones, Chief Secretary to the Treasury, said: “This Government is determined that digital transformation of the state and our public services will deliver better outcomes for people, and ensure every pound of taxpayers money is spent well.”
RELATED STORIES
Government unveils digital inclusion action plan to tackle digital poverty
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms | 5 |
After a review found that government-funded projects face overly complex spending approval processes that bog them down, the route to funding is set to be simplified.
The changes will see projects supported with larger tests if they have the potential to save money or improve public services.
Subscribe today for free
Peter Kyle, the UK’s Technology Secretary, said: “Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in Government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The research funding review comes as part of a wider push by Sir Keir Starmer’s Labour government to make the UK more attractive to tech investments, which includes creating AI Growth Zones and streamlining planning rules for projects like data centres and nuclear reactors .
“This review will help us build technology that will mean businesses can skip the admin and get on with driving growth, digital systems supporting the police are more reliable so they can keep our streets safe, and it will mean we can build new tools to speed up wait times for doctors’ appointments and get the NHS back on its feet are built,” Kyle added.
The government said four new processes will be tested from April, adding to its experimental generative AI chatbot, GOV.UK Chat.
Darren Jones, Chief Secretary to the Treasury, said: “This Government is determined that digital transformation of the state and our public services will deliver better outcomes for people, and ensure every pound of taxpayers money is spent well.”
RELATED STORIES
Government unveils digital inclusion action plan to tackle digital poverty
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms | 5 |
After a review found that government-funded projects face overly complex spending approval processes that bog them down, the route to funding is set to be simplified.
The changes will see projects supported with larger tests if they have the potential to save money or improve public services.
Subscribe today for free
Peter Kyle, the UK’s Technology Secretary, said: “Technology has immense potential to build public services that work for citizens. But a decades-old process has encouraged short-sighted thinking and outdated tech while stopping crucial innovation before it even gets going.
“These changes we’re making ensure innovation is the default. We will help give AI innovators in Government the freedom they need to chase an exciting idea and build prototypes almost immediately.”
The research funding review comes as part of a wider push by Sir Keir Starmer’s Labour government to make the UK more attractive to tech investments, which includes creating AI Growth Zones and streamlining planning rules for projects like data centres and nuclear reactors .
“This review will help us build technology that will mean businesses can skip the admin and get on with driving growth, digital systems supporting the police are more reliable so they can keep our streets safe, and it will mean we can build new tools to speed up wait times for doctors’ appointments and get the NHS back on its feet are built,” Kyle added.
The government said four new processes will be tested from April, adding to its experimental generative AI chatbot, GOV.UK Chat.
Darren Jones, Chief Secretary to the Treasury, said: “This Government is determined that digital transformation of the state and our public services will deliver better outcomes for people, and ensure every pound of taxpayers money is spent well.”
RELATED STORIES
Government unveils digital inclusion action plan to tackle digital poverty
UK Gov launches ambitious plan to make UK 'irresistible' to AI firms | 5 |
AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage.
Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to build and support increasingly high value use cases. Imagine the following chat:
Here we see that chats can be complex and span multiple topics, such as event planning, team building, and logistics. Generative AI has ushered in a two-fold paradigm shift. First, LLMs give us a new thing to measure, that is, how people interact with AI systems. Second, they give us a new way to measure those interactions, that is, they give us the capability to understand and make inferences on these interactions, at scale. The Semantic Telemetry project has created new measures to classify human-AI interactions and understand user behavior, contributing to efforts in developing new approaches for measuring generative AI (opens in new tab) across various use cases.
Semantic Telemetry is a rethink of traditional telemetry–in which data is collected for understanding systems–designed for analyzing chat-based AI. We employ an innovative data science methodology that uses a large language model (LLM) to generate meaningful categorical labels, enabling us to gain insights into chat log data.
Figure 1: Prompting an LLM to classify a conversation based on LLM generated label taxonomy
This process begins with developing a set of classifications and definitions. We create these classifications by instructing an LLM to generate a short summary of the conversation, and then iteratively prompting the LLM to generate, update, and review classification labels on a batched set of summaries. This process is outlined in the paper: TnT-LLM: Text Mining at Scale with Large Language Models. We then prompt an LLM with these generated classifiers to label new unstructured (and unlabeled) chat log data.
Description of LLM generated label taxonomy process
With this approach, we have analyzed how people interact with Copilot in Bing. In this blog, we examine insights into how people are using Copilot in Bing, including how that differs from traditional search engines. Note that all analyses were conducted on anonymous Copilot interactions containing no personal information.
Topics
To get a clear picture of how people are using Copilot in Bing, we need to first classify sessions into topical categories. To do this, we developed a topic classifier. We used the LLM classification approach described above to label the primary topic (domain) for the entire content of the chat. Although a single chat can cover multiple topics, for this analysis, we generated a single label for the primary topic of the conversation. We sampled five million anonymized Copilot in Bing chats during August and September 2024, and found that globally, 21% of all chats were about technology, with a high concentration of these chats in programming and scripting and computers and electronics.
Figure 2: Top Copilot in Bing topics based on anonymized data (August-September 2024)
Figure 3: Frequent topic summaries in Technology
Figure 4: Frequent topic summaries in Entertainment
Diving into the technology category, we find a lot of professional tasks in programming and scripting, where users request problem-specific assistance such as fixing a SQL query syntax error. In computers and electronics, we observe users getting help with tasks like adjusting screen brightness and troubleshooting internet connectivity issues. We can compare this with our second most common topic, entertainment, in which we see users seeking information related to personal activities like hiking and game nights.
We also note that top topics differ by platform. The figure below depicts topic popularity based on mobile and desktop usage. Mobile device users tend to use the chat for more personal-related tasks such as helping to plant a garden or understanding medical symptoms whereas desktop users conduct more professional tasks like revising an email.
Figure 5: Top topics for desktop users and mobile users
Spotlight: Blog post MedFuzz: Exploring the robustness of LLMs on medical challenge problems Medfuzz tests LLMs by breaking benchmark assumptions, exposing vulnerabilities to bolster real-world accuracy. Read more Opens in a new tab
Search versus Copilot
Beyond analyzing topics, we compared Copilot in Bing usage to that of traditional search. Chat extends beyond traditional online search by enabling users to summarize, generate, compare, and analyze information. Human-AI interactions are conversational and more complex than traditional search (Figure 6).
Figure 6: Bing Search Query compared to Copilot in Bing Conversation
A major differentiation between search and chat is the ability to ask more complex questions, but how can we measure this? We think of complexity as a scale ranging from simply asking chat to look up information to evaluating several ideas. We aim to understand the difficulty of a task if performed by a human without the assistance of AI. To achieve this, we developed the task complexity classifier, which assesses task difficulty using Anderson and Krathwohl’s Taxonomy of Learning Objectives (opens in new tab). For our analysis, we have grouped the learning objectives into two categories: low complexity and high complexity. Any task more complicated than information lookup is classified as high complexity. Note that this would be very challenging to classify using traditional data science techniques.
Description of task complexity and 6 categories of the Anderson and Krathwohl’s Taxonomy of Learning Objectives
Comparing low versus high complexity tasks, most chat interactions were categorized as high complexity (78.9%), meaning that they were more complex than looking up information. Programming and scripting, marketing and sales, and creative and professional writing are topics in which users engage in higher complexity tasks (Figure 7) such as learning a skill, troubleshooting a problem, or writing an article.
Figure 7: Most and least complex topics based on percentage of high complexity tasks.
Travel and tourism and history and culture scored lowest in complexity, with users looking up information like flight times and latest news updates.
Demo of task complexity and topics on anonymous Copilot interactions
When should you use chat instead of search? A 2024 Microsoft Research study: The Use of Generative Search Engines for Knowledge Work and Complex Tasks, suggests that people are seeing value in technical, complex tasks such as web development and data analysis. Bing Search contained more queries with lower complexity focused on non-professional areas, like gaming and entertainment, travel and tourism, and fashion and beauty, while chat had a greater distribution of complex technical tasks. (Figure 8).
Figure 8: Comparison of Bing Search and Copilot in Bing for anonymized sample data (May-June 2023)
Conclusion
LLMs have enabled a new era of high-quality human-AI interaction, and with it, the capability to analyze those same interactions with high fidelity, at scale, and in near real-time. We are now able to obtain actionable insight from complex data that is not possible with traditional data science pattern-matching methods. LLM-generated classifications are pushing research into new directions that will ultimately improve user experience and satisfaction when using chat and other user-AI interaction tools.
This analysis indicates that Copilot in Bing is enabling users to do more complex work, specifically in areas such as technology. In our next post, we will explore how Copilot in Bing is supporting professional knowledge work and how we can use these measures as indicators for retention and engagement.
FOOTNOTE: This research was conducted at the time the feature Copilot in Bing was available as part of the Bing service; since October 2024 Copilot in Bing has been deprecated in favor of the standalone Microsoft Copilot service.
References:
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory Into Practice, 41(4), 212–218. https://doi.org/10.1207/s15430421tip4104_2 (opens in new tab)
Opens in a new tab | 5 |
AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage.
Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to build and support increasingly high value use cases. Imagine the following chat:
Here we see that chats can be complex and span multiple topics, such as event planning, team building, and logistics. Generative AI has ushered in a two-fold paradigm shift. First, LLMs give us a new thing to measure, that is, how people interact with AI systems. Second, they give us a new way to measure those interactions, that is, they give us the capability to understand and make inferences on these interactions, at scale. The Semantic Telemetry project has created new measures to classify human-AI interactions and understand user behavior, contributing to efforts in developing new approaches for measuring generative AI (opens in new tab) across various use cases.
Semantic Telemetry is a rethink of traditional telemetry–in which data is collected for understanding systems–designed for analyzing chat-based AI. We employ an innovative data science methodology that uses a large language model (LLM) to generate meaningful categorical labels, enabling us to gain insights into chat log data.
Figure 1: Prompting an LLM to classify a conversation based on LLM generated label taxonomy
This process begins with developing a set of classifications and definitions. We create these classifications by instructing an LLM to generate a short summary of the conversation, and then iteratively prompting the LLM to generate, update, and review classification labels on a batched set of summaries. This process is outlined in the paper: TnT-LLM: Text Mining at Scale with Large Language Models. We then prompt an LLM with these generated classifiers to label new unstructured (and unlabeled) chat log data.
Description of LLM generated label taxonomy process
With this approach, we have analyzed how people interact with Copilot in Bing. In this blog, we examine insights into how people are using Copilot in Bing, including how that differs from traditional search engines. Note that all analyses were conducted on anonymous Copilot interactions containing no personal information.
Topics
To get a clear picture of how people are using Copilot in Bing, we need to first classify sessions into topical categories. To do this, we developed a topic classifier. We used the LLM classification approach described above to label the primary topic (domain) for the entire content of the chat. Although a single chat can cover multiple topics, for this analysis, we generated a single label for the primary topic of the conversation. We sampled five million anonymized Copilot in Bing chats during August and September 2024, and found that globally, 21% of all chats were about technology, with a high concentration of these chats in programming and scripting and computers and electronics.
Figure 2: Top Copilot in Bing topics based on anonymized data (August-September 2024)
Figure 3: Frequent topic summaries in Technology
Figure 4: Frequent topic summaries in Entertainment
Diving into the technology category, we find a lot of professional tasks in programming and scripting, where users request problem-specific assistance such as fixing a SQL query syntax error. In computers and electronics, we observe users getting help with tasks like adjusting screen brightness and troubleshooting internet connectivity issues. We can compare this with our second most common topic, entertainment, in which we see users seeking information related to personal activities like hiking and game nights.
We also note that top topics differ by platform. The figure below depicts topic popularity based on mobile and desktop usage. Mobile device users tend to use the chat for more personal-related tasks such as helping to plant a garden or understanding medical symptoms whereas desktop users conduct more professional tasks like revising an email.
Figure 5: Top topics for desktop users and mobile users
Spotlight: Blog post MedFuzz: Exploring the robustness of LLMs on medical challenge problems Medfuzz tests LLMs by breaking benchmark assumptions, exposing vulnerabilities to bolster real-world accuracy. Read more Opens in a new tab
Search versus Copilot
Beyond analyzing topics, we compared Copilot in Bing usage to that of traditional search. Chat extends beyond traditional online search by enabling users to summarize, generate, compare, and analyze information. Human-AI interactions are conversational and more complex than traditional search (Figure 6).
Figure 6: Bing Search Query compared to Copilot in Bing Conversation
A major differentiation between search and chat is the ability to ask more complex questions, but how can we measure this? We think of complexity as a scale ranging from simply asking chat to look up information to evaluating several ideas. We aim to understand the difficulty of a task if performed by a human without the assistance of AI. To achieve this, we developed the task complexity classifier, which assesses task difficulty using Anderson and Krathwohl’s Taxonomy of Learning Objectives (opens in new tab). For our analysis, we have grouped the learning objectives into two categories: low complexity and high complexity. Any task more complicated than information lookup is classified as high complexity. Note that this would be very challenging to classify using traditional data science techniques.
Description of task complexity and 6 categories of the Anderson and Krathwohl’s Taxonomy of Learning Objectives
Comparing low versus high complexity tasks, most chat interactions were categorized as high complexity (78.9%), meaning that they were more complex than looking up information. Programming and scripting, marketing and sales, and creative and professional writing are topics in which users engage in higher complexity tasks (Figure 7) such as learning a skill, troubleshooting a problem, or writing an article.
Figure 7: Most and least complex topics based on percentage of high complexity tasks.
Travel and tourism and history and culture scored lowest in complexity, with users looking up information like flight times and latest news updates.
Demo of task complexity and topics on anonymous Copilot interactions
When should you use chat instead of search? A 2024 Microsoft Research study: The Use of Generative Search Engines for Knowledge Work and Complex Tasks, suggests that people are seeing value in technical, complex tasks such as web development and data analysis. Bing Search contained more queries with lower complexity focused on non-professional areas, like gaming and entertainment, travel and tourism, and fashion and beauty, while chat had a greater distribution of complex technical tasks. (Figure 8).
Figure 8: Comparison of Bing Search and Copilot in Bing for anonymized sample data (May-June 2023)
Conclusion
LLMs have enabled a new era of high-quality human-AI interaction, and with it, the capability to analyze those same interactions with high fidelity, at scale, and in near real-time. We are now able to obtain actionable insight from complex data that is not possible with traditional data science pattern-matching methods. LLM-generated classifications are pushing research into new directions that will ultimately improve user experience and satisfaction when using chat and other user-AI interaction tools.
This analysis indicates that Copilot in Bing is enabling users to do more complex work, specifically in areas such as technology. In our next post, we will explore how Copilot in Bing is supporting professional knowledge work and how we can use these measures as indicators for retention and engagement.
FOOTNOTE: This research was conducted at the time the feature Copilot in Bing was available as part of the Bing service; since October 2024 Copilot in Bing has been deprecated in favor of the standalone Microsoft Copilot service.
References:
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory Into Practice, 41(4), 212–218. https://doi.org/10.1207/s15430421tip4104_2 (opens in new tab)
Opens in a new tab | 5 |
AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage.
Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to build and support increasingly high value use cases. Imagine the following chat:
Here we see that chats can be complex and span multiple topics, such as event planning, team building, and logistics. Generative AI has ushered in a two-fold paradigm shift. First, LLMs give us a new thing to measure, that is, how people interact with AI systems. Second, they give us a new way to measure those interactions, that is, they give us the capability to understand and make inferences on these interactions, at scale. The Semantic Telemetry project has created new measures to classify human-AI interactions and understand user behavior, contributing to efforts in developing new approaches for measuring generative AI (opens in new tab) across various use cases.
Semantic Telemetry is a rethink of traditional telemetry–in which data is collected for understanding systems–designed for analyzing chat-based AI. We employ an innovative data science methodology that uses a large language model (LLM) to generate meaningful categorical labels, enabling us to gain insights into chat log data.
Figure 1: Prompting an LLM to classify a conversation based on LLM generated label taxonomy
This process begins with developing a set of classifications and definitions. We create these classifications by instructing an LLM to generate a short summary of the conversation, and then iteratively prompting the LLM to generate, update, and review classification labels on a batched set of summaries. This process is outlined in the paper: TnT-LLM: Text Mining at Scale with Large Language Models. We then prompt an LLM with these generated classifiers to label new unstructured (and unlabeled) chat log data.
Description of LLM generated label taxonomy process
With this approach, we have analyzed how people interact with Copilot in Bing. In this blog, we examine insights into how people are using Copilot in Bing, including how that differs from traditional search engines. Note that all analyses were conducted on anonymous Copilot interactions containing no personal information.
Topics
To get a clear picture of how people are using Copilot in Bing, we need to first classify sessions into topical categories. To do this, we developed a topic classifier. We used the LLM classification approach described above to label the primary topic (domain) for the entire content of the chat. Although a single chat can cover multiple topics, for this analysis, we generated a single label for the primary topic of the conversation. We sampled five million anonymized Copilot in Bing chats during August and September 2024, and found that globally, 21% of all chats were about technology, with a high concentration of these chats in programming and scripting and computers and electronics.
Figure 2: Top Copilot in Bing topics based on anonymized data (August-September 2024)
Figure 3: Frequent topic summaries in Technology
Figure 4: Frequent topic summaries in Entertainment
Diving into the technology category, we find a lot of professional tasks in programming and scripting, where users request problem-specific assistance such as fixing a SQL query syntax error. In computers and electronics, we observe users getting help with tasks like adjusting screen brightness and troubleshooting internet connectivity issues. We can compare this with our second most common topic, entertainment, in which we see users seeking information related to personal activities like hiking and game nights.
We also note that top topics differ by platform. The figure below depicts topic popularity based on mobile and desktop usage. Mobile device users tend to use the chat for more personal-related tasks such as helping to plant a garden or understanding medical symptoms whereas desktop users conduct more professional tasks like revising an email.
Figure 5: Top topics for desktop users and mobile users
Spotlight: Blog post MedFuzz: Exploring the robustness of LLMs on medical challenge problems Medfuzz tests LLMs by breaking benchmark assumptions, exposing vulnerabilities to bolster real-world accuracy. Read more Opens in a new tab
Search versus Copilot
Beyond analyzing topics, we compared Copilot in Bing usage to that of traditional search. Chat extends beyond traditional online search by enabling users to summarize, generate, compare, and analyze information. Human-AI interactions are conversational and more complex than traditional search (Figure 6).
Figure 6: Bing Search Query compared to Copilot in Bing Conversation
A major differentiation between search and chat is the ability to ask more complex questions, but how can we measure this? We think of complexity as a scale ranging from simply asking chat to look up information to evaluating several ideas. We aim to understand the difficulty of a task if performed by a human without the assistance of AI. To achieve this, we developed the task complexity classifier, which assesses task difficulty using Anderson and Krathwohl’s Taxonomy of Learning Objectives (opens in new tab). For our analysis, we have grouped the learning objectives into two categories: low complexity and high complexity. Any task more complicated than information lookup is classified as high complexity. Note that this would be very challenging to classify using traditional data science techniques.
Description of task complexity and 6 categories of the Anderson and Krathwohl’s Taxonomy of Learning Objectives
Comparing low versus high complexity tasks, most chat interactions were categorized as high complexity (78.9%), meaning that they were more complex than looking up information. Programming and scripting, marketing and sales, and creative and professional writing are topics in which users engage in higher complexity tasks (Figure 7) such as learning a skill, troubleshooting a problem, or writing an article.
Figure 7: Most and least complex topics based on percentage of high complexity tasks.
Travel and tourism and history and culture scored lowest in complexity, with users looking up information like flight times and latest news updates.
Demo of task complexity and topics on anonymous Copilot interactions
When should you use chat instead of search? A 2024 Microsoft Research study: The Use of Generative Search Engines for Knowledge Work and Complex Tasks, suggests that people are seeing value in technical, complex tasks such as web development and data analysis. Bing Search contained more queries with lower complexity focused on non-professional areas, like gaming and entertainment, travel and tourism, and fashion and beauty, while chat had a greater distribution of complex technical tasks. (Figure 8).
Figure 8: Comparison of Bing Search and Copilot in Bing for anonymized sample data (May-June 2023)
Conclusion
LLMs have enabled a new era of high-quality human-AI interaction, and with it, the capability to analyze those same interactions with high fidelity, at scale, and in near real-time. We are now able to obtain actionable insight from complex data that is not possible with traditional data science pattern-matching methods. LLM-generated classifications are pushing research into new directions that will ultimately improve user experience and satisfaction when using chat and other user-AI interaction tools.
This analysis indicates that Copilot in Bing is enabling users to do more complex work, specifically in areas such as technology. In our next post, we will explore how Copilot in Bing is supporting professional knowledge work and how we can use these measures as indicators for retention and engagement.
FOOTNOTE: This research was conducted at the time the feature Copilot in Bing was available as part of the Bing service; since October 2024 Copilot in Bing has been deprecated in favor of the standalone Microsoft Copilot service.
References:
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory Into Practice, 41(4), 212–218. https://doi.org/10.1207/s15430421tip4104_2 (opens in new tab)
Opens in a new tab | 5 |
AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage.
Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to build and support increasingly high value use cases. Imagine the following chat:
Here we see that chats can be complex and span multiple topics, such as event planning, team building, and logistics. Generative AI has ushered in a two-fold paradigm shift. First, LLMs give us a new thing to measure, that is, how people interact with AI systems. Second, they give us a new way to measure those interactions, that is, they give us the capability to understand and make inferences on these interactions, at scale. The Semantic Telemetry project has created new measures to classify human-AI interactions and understand user behavior, contributing to efforts in developing new approaches for measuring generative AI (opens in new tab) across various use cases.
Semantic Telemetry is a rethink of traditional telemetry–in which data is collected for understanding systems–designed for analyzing chat-based AI. We employ an innovative data science methodology that uses a large language model (LLM) to generate meaningful categorical labels, enabling us to gain insights into chat log data.
Figure 1: Prompting an LLM to classify a conversation based on LLM generated label taxonomy
This process begins with developing a set of classifications and definitions. We create these classifications by instructing an LLM to generate a short summary of the conversation, and then iteratively prompting the LLM to generate, update, and review classification labels on a batched set of summaries. This process is outlined in the paper: TnT-LLM: Text Mining at Scale with Large Language Models. We then prompt an LLM with these generated classifiers to label new unstructured (and unlabeled) chat log data.
Description of LLM generated label taxonomy process
With this approach, we have analyzed how people interact with Copilot in Bing. In this blog, we examine insights into how people are using Copilot in Bing, including how that differs from traditional search engines. Note that all analyses were conducted on anonymous Copilot interactions containing no personal information.
Topics
To get a clear picture of how people are using Copilot in Bing, we need to first classify sessions into topical categories. To do this, we developed a topic classifier. We used the LLM classification approach described above to label the primary topic (domain) for the entire content of the chat. Although a single chat can cover multiple topics, for this analysis, we generated a single label for the primary topic of the conversation. We sampled five million anonymized Copilot in Bing chats during August and September 2024, and found that globally, 21% of all chats were about technology, with a high concentration of these chats in programming and scripting and computers and electronics.
Figure 2: Top Copilot in Bing topics based on anonymized data (August-September 2024)
Figure 3: Frequent topic summaries in Technology
Figure 4: Frequent topic summaries in Entertainment
Diving into the technology category, we find a lot of professional tasks in programming and scripting, where users request problem-specific assistance such as fixing a SQL query syntax error. In computers and electronics, we observe users getting help with tasks like adjusting screen brightness and troubleshooting internet connectivity issues. We can compare this with our second most common topic, entertainment, in which we see users seeking information related to personal activities like hiking and game nights.
We also note that top topics differ by platform. The figure below depicts topic popularity based on mobile and desktop usage. Mobile device users tend to use the chat for more personal-related tasks such as helping to plant a garden or understanding medical symptoms whereas desktop users conduct more professional tasks like revising an email.
Figure 5: Top topics for desktop users and mobile users
Spotlight: Blog post MedFuzz: Exploring the robustness of LLMs on medical challenge problems Medfuzz tests LLMs by breaking benchmark assumptions, exposing vulnerabilities to bolster real-world accuracy. Read more Opens in a new tab
Search versus Copilot
Beyond analyzing topics, we compared Copilot in Bing usage to that of traditional search. Chat extends beyond traditional online search by enabling users to summarize, generate, compare, and analyze information. Human-AI interactions are conversational and more complex than traditional search (Figure 6).
Figure 6: Bing Search Query compared to Copilot in Bing Conversation
A major differentiation between search and chat is the ability to ask more complex questions, but how can we measure this? We think of complexity as a scale ranging from simply asking chat to look up information to evaluating several ideas. We aim to understand the difficulty of a task if performed by a human without the assistance of AI. To achieve this, we developed the task complexity classifier, which assesses task difficulty using Anderson and Krathwohl’s Taxonomy of Learning Objectives (opens in new tab). For our analysis, we have grouped the learning objectives into two categories: low complexity and high complexity. Any task more complicated than information lookup is classified as high complexity. Note that this would be very challenging to classify using traditional data science techniques.
Description of task complexity and 6 categories of the Anderson and Krathwohl’s Taxonomy of Learning Objectives
Comparing low versus high complexity tasks, most chat interactions were categorized as high complexity (78.9%), meaning that they were more complex than looking up information. Programming and scripting, marketing and sales, and creative and professional writing are topics in which users engage in higher complexity tasks (Figure 7) such as learning a skill, troubleshooting a problem, or writing an article.
Figure 7: Most and least complex topics based on percentage of high complexity tasks.
Travel and tourism and history and culture scored lowest in complexity, with users looking up information like flight times and latest news updates.
Demo of task complexity and topics on anonymous Copilot interactions
When should you use chat instead of search? A 2024 Microsoft Research study: The Use of Generative Search Engines for Knowledge Work and Complex Tasks, suggests that people are seeing value in technical, complex tasks such as web development and data analysis. Bing Search contained more queries with lower complexity focused on non-professional areas, like gaming and entertainment, travel and tourism, and fashion and beauty, while chat had a greater distribution of complex technical tasks. (Figure 8).
Figure 8: Comparison of Bing Search and Copilot in Bing for anonymized sample data (May-June 2023)
Conclusion
LLMs have enabled a new era of high-quality human-AI interaction, and with it, the capability to analyze those same interactions with high fidelity, at scale, and in near real-time. We are now able to obtain actionable insight from complex data that is not possible with traditional data science pattern-matching methods. LLM-generated classifications are pushing research into new directions that will ultimately improve user experience and satisfaction when using chat and other user-AI interaction tools.
This analysis indicates that Copilot in Bing is enabling users to do more complex work, specifically in areas such as technology. In our next post, we will explore how Copilot in Bing is supporting professional knowledge work and how we can use these measures as indicators for retention and engagement.
FOOTNOTE: This research was conducted at the time the feature Copilot in Bing was available as part of the Bing service; since October 2024 Copilot in Bing has been deprecated in favor of the standalone Microsoft Copilot service.
References:
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory Into Practice, 41(4), 212–218. https://doi.org/10.1207/s15430421tip4104_2 (opens in new tab)
Opens in a new tab | 5 |
AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage.
Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to build and support increasingly high value use cases. Imagine the following chat:
Here we see that chats can be complex and span multiple topics, such as event planning, team building, and logistics. Generative AI has ushered in a two-fold paradigm shift. First, LLMs give us a new thing to measure, that is, how people interact with AI systems. Second, they give us a new way to measure those interactions, that is, they give us the capability to understand and make inferences on these interactions, at scale. The Semantic Telemetry project has created new measures to classify human-AI interactions and understand user behavior, contributing to efforts in developing new approaches for measuring generative AI (opens in new tab) across various use cases.
Semantic Telemetry is a rethink of traditional telemetry–in which data is collected for understanding systems–designed for analyzing chat-based AI. We employ an innovative data science methodology that uses a large language model (LLM) to generate meaningful categorical labels, enabling us to gain insights into chat log data.
Figure 1: Prompting an LLM to classify a conversation based on LLM generated label taxonomy
This process begins with developing a set of classifications and definitions. We create these classifications by instructing an LLM to generate a short summary of the conversation, and then iteratively prompting the LLM to generate, update, and review classification labels on a batched set of summaries. This process is outlined in the paper: TnT-LLM: Text Mining at Scale with Large Language Models. We then prompt an LLM with these generated classifiers to label new unstructured (and unlabeled) chat log data.
Description of LLM generated label taxonomy process
With this approach, we have analyzed how people interact with Copilot in Bing. In this blog, we examine insights into how people are using Copilot in Bing, including how that differs from traditional search engines. Note that all analyses were conducted on anonymous Copilot interactions containing no personal information.
Topics
To get a clear picture of how people are using Copilot in Bing, we need to first classify sessions into topical categories. To do this, we developed a topic classifier. We used the LLM classification approach described above to label the primary topic (domain) for the entire content of the chat. Although a single chat can cover multiple topics, for this analysis, we generated a single label for the primary topic of the conversation. We sampled five million anonymized Copilot in Bing chats during August and September 2024, and found that globally, 21% of all chats were about technology, with a high concentration of these chats in programming and scripting and computers and electronics.
Figure 2: Top Copilot in Bing topics based on anonymized data (August-September 2024)
Figure 3: Frequent topic summaries in Technology
Figure 4: Frequent topic summaries in Entertainment
Diving into the technology category, we find a lot of professional tasks in programming and scripting, where users request problem-specific assistance such as fixing a SQL query syntax error. In computers and electronics, we observe users getting help with tasks like adjusting screen brightness and troubleshooting internet connectivity issues. We can compare this with our second most common topic, entertainment, in which we see users seeking information related to personal activities like hiking and game nights.
We also note that top topics differ by platform. The figure below depicts topic popularity based on mobile and desktop usage. Mobile device users tend to use the chat for more personal-related tasks such as helping to plant a garden or understanding medical symptoms whereas desktop users conduct more professional tasks like revising an email.
Figure 5: Top topics for desktop users and mobile users
Spotlight: Blog post MedFuzz: Exploring the robustness of LLMs on medical challenge problems Medfuzz tests LLMs by breaking benchmark assumptions, exposing vulnerabilities to bolster real-world accuracy. Read more Opens in a new tab
Search versus Copilot
Beyond analyzing topics, we compared Copilot in Bing usage to that of traditional search. Chat extends beyond traditional online search by enabling users to summarize, generate, compare, and analyze information. Human-AI interactions are conversational and more complex than traditional search (Figure 6).
Figure 6: Bing Search Query compared to Copilot in Bing Conversation
A major differentiation between search and chat is the ability to ask more complex questions, but how can we measure this? We think of complexity as a scale ranging from simply asking chat to look up information to evaluating several ideas. We aim to understand the difficulty of a task if performed by a human without the assistance of AI. To achieve this, we developed the task complexity classifier, which assesses task difficulty using Anderson and Krathwohl’s Taxonomy of Learning Objectives (opens in new tab). For our analysis, we have grouped the learning objectives into two categories: low complexity and high complexity. Any task more complicated than information lookup is classified as high complexity. Note that this would be very challenging to classify using traditional data science techniques.
Description of task complexity and 6 categories of the Anderson and Krathwohl’s Taxonomy of Learning Objectives
Comparing low versus high complexity tasks, most chat interactions were categorized as high complexity (78.9%), meaning that they were more complex than looking up information. Programming and scripting, marketing and sales, and creative and professional writing are topics in which users engage in higher complexity tasks (Figure 7) such as learning a skill, troubleshooting a problem, or writing an article.
Figure 7: Most and least complex topics based on percentage of high complexity tasks.
Travel and tourism and history and culture scored lowest in complexity, with users looking up information like flight times and latest news updates.
Demo of task complexity and topics on anonymous Copilot interactions
When should you use chat instead of search? A 2024 Microsoft Research study: The Use of Generative Search Engines for Knowledge Work and Complex Tasks, suggests that people are seeing value in technical, complex tasks such as web development and data analysis. Bing Search contained more queries with lower complexity focused on non-professional areas, like gaming and entertainment, travel and tourism, and fashion and beauty, while chat had a greater distribution of complex technical tasks. (Figure 8).
Figure 8: Comparison of Bing Search and Copilot in Bing for anonymized sample data (May-June 2023)
Conclusion
LLMs have enabled a new era of high-quality human-AI interaction, and with it, the capability to analyze those same interactions with high fidelity, at scale, and in near real-time. We are now able to obtain actionable insight from complex data that is not possible with traditional data science pattern-matching methods. LLM-generated classifications are pushing research into new directions that will ultimately improve user experience and satisfaction when using chat and other user-AI interaction tools.
This analysis indicates that Copilot in Bing is enabling users to do more complex work, specifically in areas such as technology. In our next post, we will explore how Copilot in Bing is supporting professional knowledge work and how we can use these measures as indicators for retention and engagement.
FOOTNOTE: This research was conducted at the time the feature Copilot in Bing was available as part of the Bing service; since October 2024 Copilot in Bing has been deprecated in favor of the standalone Microsoft Copilot service.
References:
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory Into Practice, 41(4), 212–218. https://doi.org/10.1207/s15430421tip4104_2 (opens in new tab)
Opens in a new tab | 5 |
ARLINGTON, Va., March 10, 2025 /PRNewswire/ -- Today Bloomberg Tax & Accounting announced the launch of two new generative AI-powered features, Bloomberg Tax Answers and AI Assistant. Bloomberg Tax Answers replaces hours of tax research with fast and precise answers to tax questions with supporting primary sources and industry leading expert analysis, enabling tax practitioners to quickly find, validate and apply information to their workflows. AI Assistant is a chat-based research tool that allows users to ask targeted questions to identify specific information from within a document, including Bloomberg Tax's market leading Portfolios, and to build a chart comparing tax information across jurisdictions. Bloomberg Tax Answers and AI Assistant are available within the Bloomberg Tax platform at no additional charge.
Bloomberg Tax Answers and AI Assistant
Bloomberg Tax Answers leverages generative AI and authoritative content to provide a brief but meaningful answer to a user's search directly on top of regular search results, with no need to learn a new tool. Each answer generated includes citations and links to the Bloomberg Tax authorities and source documents used to generate it, including select primary and secondary sources such as the Internal Revenue Code, federal and state tax agency documents, state tax statutes and regulations, and Bloomberg Tax content.
Bloomberg Tax's AI Assistant supports a number of important research tasks. Currently, the tool allows customers to generate summaries of Bloomberg Tax Analysis, saving significant time in their research by providing clear, easy-to-read overviews of documents. Users can also ask the AI Assistant specific questions about the document to quickly identify the information they are looking for. Similarly, the assistant allows customers to ask the AI Assistant to create a chart comparing tax information across jurisdictions.
"The latest AI-powered features for Bloomberg Tax & Accounting showcase our dedication to innovation and solving complex tax research challenges," said Evan Croen, head of Bloomberg Tax. "Bloomberg Tax Answers and AI Assistant deliver rapid, accurate answers and facilitate cross-jurisdictional comparisons. Additionally, users can verify information with direct access to cited source documents, enhancing reliability and trustworthiness."
"Bloomberg Tax Answers is a very efficient method in obtaining best applicable searches from a pertinent database, as opposed to scrolling past many non-relevant searches," said a senior manager at a public corporation.
Bloomberg Tax Answers and AI Assistant will be subject to ongoing refinement based on customer feedback. AI Assistant will be updated with additional research skills in the coming months.
For more information about Bloomberg Tax & Accounting's AI innovations and approach to AI, please visit https://aboutbtax.com/bhsy.
About Bloomberg Tax & Accounting
Bloomberg Tax & Accounting provides practitioner-driven research and technology solutions that deliver timely, strategic insights to enable smarter decisions. From our unparalleled Tax Management Portfolios to technology designed to streamline the most complex planning and compliance scenarios, we deliver essential news and analysis, practical perspectives, and software that help tax and accounting professionals around the globe mitigate risk and maximize business results. For more information, visit Bloomberg Tax.
SOURCE Bloomberg Tax & Accounting | 5 |
ARLINGTON, Va., March 10, 2025 /PRNewswire/ -- Today Bloomberg Tax & Accounting announced the launch of two new generative AI-powered features, Bloomberg Tax Answers and AI Assistant. Bloomberg Tax Answers replaces hours of tax research with fast and precise answers to tax questions with supporting primary sources and industry leading expert analysis, enabling tax practitioners to quickly find, validate and apply information to their workflows. AI Assistant is a chat-based research tool that allows users to ask targeted questions to identify specific information from within a document, including Bloomberg Tax's market leading Portfolios, and to build a chart comparing tax information across jurisdictions. Bloomberg Tax Answers and AI Assistant are available within the Bloomberg Tax platform at no additional charge.
Bloomberg Tax Answers and AI Assistant
Bloomberg Tax Answers leverages generative AI and authoritative content to provide a brief but meaningful answer to a user's search directly on top of regular search results, with no need to learn a new tool. Each answer generated includes citations and links to the Bloomberg Tax authorities and source documents used to generate it, including select primary and secondary sources such as the Internal Revenue Code, federal and state tax agency documents, state tax statutes and regulations, and Bloomberg Tax content.
Bloomberg Tax's AI Assistant supports a number of important research tasks. Currently, the tool allows customers to generate summaries of Bloomberg Tax Analysis, saving significant time in their research by providing clear, easy-to-read overviews of documents. Users can also ask the AI Assistant specific questions about the document to quickly identify the information they are looking for. Similarly, the assistant allows customers to ask the AI Assistant to create a chart comparing tax information across jurisdictions.
"The latest AI-powered features for Bloomberg Tax & Accounting showcase our dedication to innovation and solving complex tax research challenges," said Evan Croen, head of Bloomberg Tax. "Bloomberg Tax Answers and AI Assistant deliver rapid, accurate answers and facilitate cross-jurisdictional comparisons. Additionally, users can verify information with direct access to cited source documents, enhancing reliability and trustworthiness."
"Bloomberg Tax Answers is a very efficient method in obtaining best applicable searches from a pertinent database, as opposed to scrolling past many non-relevant searches," said a senior manager at a public corporation.
Bloomberg Tax Answers and AI Assistant will be subject to ongoing refinement based on customer feedback. AI Assistant will be updated with additional research skills in the coming months.
For more information about Bloomberg Tax & Accounting's AI innovations and approach to AI, please visit https://aboutbtax.com/bhsy.
About Bloomberg Tax & Accounting
Bloomberg Tax & Accounting provides practitioner-driven research and technology solutions that deliver timely, strategic insights to enable smarter decisions. From our unparalleled Tax Management Portfolios to technology designed to streamline the most complex planning and compliance scenarios, we deliver essential news and analysis, practical perspectives, and software that help tax and accounting professionals around the globe mitigate risk and maximize business results. For more information, visit Bloomberg Tax.
SOURCE Bloomberg Tax & Accounting | 5 |
ARLINGTON, Va., March 10, 2025 /PRNewswire/ -- Today Bloomberg Tax & Accounting announced the launch of two new generative AI-powered features, Bloomberg Tax Answers and AI Assistant. Bloomberg Tax Answers replaces hours of tax research with fast and precise answers to tax questions with supporting primary sources and industry leading expert analysis, enabling tax practitioners to quickly find, validate and apply information to their workflows. AI Assistant is a chat-based research tool that allows users to ask targeted questions to identify specific information from within a document, including Bloomberg Tax's market leading Portfolios, and to build a chart comparing tax information across jurisdictions. Bloomberg Tax Answers and AI Assistant are available within the Bloomberg Tax platform at no additional charge.
Bloomberg Tax Answers and AI Assistant
Bloomberg Tax Answers leverages generative AI and authoritative content to provide a brief but meaningful answer to a user's search directly on top of regular search results, with no need to learn a new tool. Each answer generated includes citations and links to the Bloomberg Tax authorities and source documents used to generate it, including select primary and secondary sources such as the Internal Revenue Code, federal and state tax agency documents, state tax statutes and regulations, and Bloomberg Tax content.
Bloomberg Tax's AI Assistant supports a number of important research tasks. Currently, the tool allows customers to generate summaries of Bloomberg Tax Analysis, saving significant time in their research by providing clear, easy-to-read overviews of documents. Users can also ask the AI Assistant specific questions about the document to quickly identify the information they are looking for. Similarly, the assistant allows customers to ask the AI Assistant to create a chart comparing tax information across jurisdictions.
"The latest AI-powered features for Bloomberg Tax & Accounting showcase our dedication to innovation and solving complex tax research challenges," said Evan Croen, head of Bloomberg Tax. "Bloomberg Tax Answers and AI Assistant deliver rapid, accurate answers and facilitate cross-jurisdictional comparisons. Additionally, users can verify information with direct access to cited source documents, enhancing reliability and trustworthiness."
"Bloomberg Tax Answers is a very efficient method in obtaining best applicable searches from a pertinent database, as opposed to scrolling past many non-relevant searches," said a senior manager at a public corporation.
Bloomberg Tax Answers and AI Assistant will be subject to ongoing refinement based on customer feedback. AI Assistant will be updated with additional research skills in the coming months.
For more information about Bloomberg Tax & Accounting's AI innovations and approach to AI, please visit https://aboutbtax.com/bhsy.
About Bloomberg Tax & Accounting
Bloomberg Tax & Accounting provides practitioner-driven research and technology solutions that deliver timely, strategic insights to enable smarter decisions. From our unparalleled Tax Management Portfolios to technology designed to streamline the most complex planning and compliance scenarios, we deliver essential news and analysis, practical perspectives, and software that help tax and accounting professionals around the globe mitigate risk and maximize business results. For more information, visit Bloomberg Tax.
SOURCE Bloomberg Tax & Accounting | 5 |
ARLINGTON, Va., March 10, 2025 /PRNewswire/ -- Today Bloomberg Tax & Accounting announced the launch of two new generative AI-powered features, Bloomberg Tax Answers and AI Assistant. Bloomberg Tax Answers replaces hours of tax research with fast and precise answers to tax questions with supporting primary sources and industry leading expert analysis, enabling tax practitioners to quickly find, validate and apply information to their workflows. AI Assistant is a chat-based research tool that allows users to ask targeted questions to identify specific information from within a document, including Bloomberg Tax's market leading Portfolios, and to build a chart comparing tax information across jurisdictions. Bloomberg Tax Answers and AI Assistant are available within the Bloomberg Tax platform at no additional charge.
Bloomberg Tax Answers and AI Assistant
Bloomberg Tax Answers leverages generative AI and authoritative content to provide a brief but meaningful answer to a user's search directly on top of regular search results, with no need to learn a new tool. Each answer generated includes citations and links to the Bloomberg Tax authorities and source documents used to generate it, including select primary and secondary sources such as the Internal Revenue Code, federal and state tax agency documents, state tax statutes and regulations, and Bloomberg Tax content.
Bloomberg Tax's AI Assistant supports a number of important research tasks. Currently, the tool allows customers to generate summaries of Bloomberg Tax Analysis, saving significant time in their research by providing clear, easy-to-read overviews of documents. Users can also ask the AI Assistant specific questions about the document to quickly identify the information they are looking for. Similarly, the assistant allows customers to ask the AI Assistant to create a chart comparing tax information across jurisdictions.
"The latest AI-powered features for Bloomberg Tax & Accounting showcase our dedication to innovation and solving complex tax research challenges," said Evan Croen, head of Bloomberg Tax. "Bloomberg Tax Answers and AI Assistant deliver rapid, accurate answers and facilitate cross-jurisdictional comparisons. Additionally, users can verify information with direct access to cited source documents, enhancing reliability and trustworthiness."
"Bloomberg Tax Answers is a very efficient method in obtaining best applicable searches from a pertinent database, as opposed to scrolling past many non-relevant searches," said a senior manager at a public corporation.
Bloomberg Tax Answers and AI Assistant will be subject to ongoing refinement based on customer feedback. AI Assistant will be updated with additional research skills in the coming months.
For more information about Bloomberg Tax & Accounting's AI innovations and approach to AI, please visit https://aboutbtax.com/bhsy.
About Bloomberg Tax & Accounting
Bloomberg Tax & Accounting provides practitioner-driven research and technology solutions that deliver timely, strategic insights to enable smarter decisions. From our unparalleled Tax Management Portfolios to technology designed to streamline the most complex planning and compliance scenarios, we deliver essential news and analysis, practical perspectives, and software that help tax and accounting professionals around the globe mitigate risk and maximize business results. For more information, visit Bloomberg Tax.
SOURCE Bloomberg Tax & Accounting | 5 |
ARLINGTON, Va., March 10, 2025 /PRNewswire/ -- Today Bloomberg Tax & Accounting announced the launch of two new generative AI-powered features, Bloomberg Tax Answers and AI Assistant. Bloomberg Tax Answers replaces hours of tax research with fast and precise answers to tax questions with supporting primary sources and industry leading expert analysis, enabling tax practitioners to quickly find, validate and apply information to their workflows. AI Assistant is a chat-based research tool that allows users to ask targeted questions to identify specific information from within a document, including Bloomberg Tax's market leading Portfolios, and to build a chart comparing tax information across jurisdictions. Bloomberg Tax Answers and AI Assistant are available within the Bloomberg Tax platform at no additional charge.
Bloomberg Tax Answers and AI Assistant
Bloomberg Tax Answers leverages generative AI and authoritative content to provide a brief but meaningful answer to a user's search directly on top of regular search results, with no need to learn a new tool. Each answer generated includes citations and links to the Bloomberg Tax authorities and source documents used to generate it, including select primary and secondary sources such as the Internal Revenue Code, federal and state tax agency documents, state tax statutes and regulations, and Bloomberg Tax content.
Bloomberg Tax's AI Assistant supports a number of important research tasks. Currently, the tool allows customers to generate summaries of Bloomberg Tax Analysis, saving significant time in their research by providing clear, easy-to-read overviews of documents. Users can also ask the AI Assistant specific questions about the document to quickly identify the information they are looking for. Similarly, the assistant allows customers to ask the AI Assistant to create a chart comparing tax information across jurisdictions.
"The latest AI-powered features for Bloomberg Tax & Accounting showcase our dedication to innovation and solving complex tax research challenges," said Evan Croen, head of Bloomberg Tax. "Bloomberg Tax Answers and AI Assistant deliver rapid, accurate answers and facilitate cross-jurisdictional comparisons. Additionally, users can verify information with direct access to cited source documents, enhancing reliability and trustworthiness."
"Bloomberg Tax Answers is a very efficient method in obtaining best applicable searches from a pertinent database, as opposed to scrolling past many non-relevant searches," said a senior manager at a public corporation.
Bloomberg Tax Answers and AI Assistant will be subject to ongoing refinement based on customer feedback. AI Assistant will be updated with additional research skills in the coming months.
For more information about Bloomberg Tax & Accounting's AI innovations and approach to AI, please visit https://aboutbtax.com/bhsy.
About Bloomberg Tax & Accounting
Bloomberg Tax & Accounting provides practitioner-driven research and technology solutions that deliver timely, strategic insights to enable smarter decisions. From our unparalleled Tax Management Portfolios to technology designed to streamline the most complex planning and compliance scenarios, we deliver essential news and analysis, practical perspectives, and software that help tax and accounting professionals around the globe mitigate risk and maximize business results. For more information, visit Bloomberg Tax.
SOURCE Bloomberg Tax & Accounting | 5 |
The Royal Swedish Academy of Sciences announced the prize in Stockholm.
An American professor and a British-Canadian professor won the Nobel Prize in Physics on Tuesday for their decadeslong, trailblazing research forming the building blocks of artificial intelligence.
John J. Hopfield, 91, was awarded the honor alongside Geoffrey E. Hinton, 76, who left his job at Google last year so he could speak freely about his concerns over the technology.
The pair are central figures in the creation of modern-day AI.
Since the 1980s, they have been using tools from physics to develop the foundations of what is known as “machine learning,” in which computers are fed masses of data to learn an array of tasks — from diagnosing diseases to knowing people’s favorite streaming shows.
Their research “formed the building blocks of machine learning, that can aid humans in making faster and more reliable decisions,” Ellen Moons, chair of the Nobel Committee for Physics, told a news conference.
The use of this technology has “become part of our daily lives, for instance in facial recognition and language translation,” Moons said, while warning that AI’s “rapid development has also raised concerns about our future.”
The machine-learning revolution can arguably be traced back to Hopfield, a Chicago-born emeritus professor at Princeton University.
Physicist, molecular biologist and neuroscientist John J. Hopfield.
In 1982, he invented the “Hopfield network,” a type of neural network — as these machine-learning programs are known — that was capable of mirroring certain functions of the human brain and recalling “memories” using only partial information.
Hinton is a British-Canadian professor at the University of Toronto who is often referred to as one of the “godfathers of AI.” He used Hopfield’s invention to come up with his own network capable of recognizing shared characteristics among large sets of data. An everyday use for this might be classifying lots of images based on things contained within them.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection,” Hinton said Tuesday, quoted by the Royal Swedish Academy of Sciences that announced the prize. “I was going to have an MRI scan today but I’ll have to cancel that!”
He worked for a decade at Google, becoming one of the world’s most renowned voices on AI. He very publicly quit his job last May, posting on X that he made the decision “so that I could talk about the dangers of AI.”
“It is hard to see how you can prevent the bad actors from using it for bad things,” Hinton said in an interview with The New York Times. | 5 |
The Royal Swedish Academy of Sciences announced the prize in Stockholm.
An American professor and a British-Canadian professor won the Nobel Prize in Physics on Tuesday for their decadeslong, trailblazing research forming the building blocks of artificial intelligence.
John J. Hopfield, 91, was awarded the honor alongside Geoffrey E. Hinton, 76, who left his job at Google last year so he could speak freely about his concerns over the technology.
The pair are central figures in the creation of modern-day AI.
Since the 1980s, they have been using tools from physics to develop the foundations of what is known as “machine learning,” in which computers are fed masses of data to learn an array of tasks — from diagnosing diseases to knowing people’s favorite streaming shows.
Their research “formed the building blocks of machine learning, that can aid humans in making faster and more reliable decisions,” Ellen Moons, chair of the Nobel Committee for Physics, told a news conference.
The use of this technology has “become part of our daily lives, for instance in facial recognition and language translation,” Moons said, while warning that AI’s “rapid development has also raised concerns about our future.”
The machine-learning revolution can arguably be traced back to Hopfield, a Chicago-born emeritus professor at Princeton University.
Physicist, molecular biologist and neuroscientist John J. Hopfield.
In 1982, he invented the “Hopfield network,” a type of neural network — as these machine-learning programs are known — that was capable of mirroring certain functions of the human brain and recalling “memories” using only partial information.
Hinton is a British-Canadian professor at the University of Toronto who is often referred to as one of the “godfathers of AI.” He used Hopfield’s invention to come up with his own network capable of recognizing shared characteristics among large sets of data. An everyday use for this might be classifying lots of images based on things contained within them.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection,” Hinton said Tuesday, quoted by the Royal Swedish Academy of Sciences that announced the prize. “I was going to have an MRI scan today but I’ll have to cancel that!”
He worked for a decade at Google, becoming one of the world’s most renowned voices on AI. He very publicly quit his job last May, posting on X that he made the decision “so that I could talk about the dangers of AI.”
“It is hard to see how you can prevent the bad actors from using it for bad things,” Hinton said in an interview with The New York Times. | 5 |
The Royal Swedish Academy of Sciences announced the prize in Stockholm.
An American professor and a British-Canadian professor won the Nobel Prize in Physics on Tuesday for their decadeslong, trailblazing research forming the building blocks of artificial intelligence.
John J. Hopfield, 91, was awarded the honor alongside Geoffrey E. Hinton, 76, who left his job at Google last year so he could speak freely about his concerns over the technology.
The pair are central figures in the creation of modern-day AI.
Since the 1980s, they have been using tools from physics to develop the foundations of what is known as “machine learning,” in which computers are fed masses of data to learn an array of tasks — from diagnosing diseases to knowing people’s favorite streaming shows.
Their research “formed the building blocks of machine learning, that can aid humans in making faster and more reliable decisions,” Ellen Moons, chair of the Nobel Committee for Physics, told a news conference.
The use of this technology has “become part of our daily lives, for instance in facial recognition and language translation,” Moons said, while warning that AI’s “rapid development has also raised concerns about our future.”
The machine-learning revolution can arguably be traced back to Hopfield, a Chicago-born emeritus professor at Princeton University.
Physicist, molecular biologist and neuroscientist John J. Hopfield.
In 1982, he invented the “Hopfield network,” a type of neural network — as these machine-learning programs are known — that was capable of mirroring certain functions of the human brain and recalling “memories” using only partial information.
Hinton is a British-Canadian professor at the University of Toronto who is often referred to as one of the “godfathers of AI.” He used Hopfield’s invention to come up with his own network capable of recognizing shared characteristics among large sets of data. An everyday use for this might be classifying lots of images based on things contained within them.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection,” Hinton said Tuesday, quoted by the Royal Swedish Academy of Sciences that announced the prize. “I was going to have an MRI scan today but I’ll have to cancel that!”
He worked for a decade at Google, becoming one of the world’s most renowned voices on AI. He very publicly quit his job last May, posting on X that he made the decision “so that I could talk about the dangers of AI.”
“It is hard to see how you can prevent the bad actors from using it for bad things,” Hinton said in an interview with The New York Times. | 5 |
The Royal Swedish Academy of Sciences announced the prize in Stockholm.
An American professor and a British-Canadian professor won the Nobel Prize in Physics on Tuesday for their decadeslong, trailblazing research forming the building blocks of artificial intelligence.
John J. Hopfield, 91, was awarded the honor alongside Geoffrey E. Hinton, 76, who left his job at Google last year so he could speak freely about his concerns over the technology.
The pair are central figures in the creation of modern-day AI.
Since the 1980s, they have been using tools from physics to develop the foundations of what is known as “machine learning,” in which computers are fed masses of data to learn an array of tasks — from diagnosing diseases to knowing people’s favorite streaming shows.
Their research “formed the building blocks of machine learning, that can aid humans in making faster and more reliable decisions,” Ellen Moons, chair of the Nobel Committee for Physics, told a news conference.
The use of this technology has “become part of our daily lives, for instance in facial recognition and language translation,” Moons said, while warning that AI’s “rapid development has also raised concerns about our future.”
The machine-learning revolution can arguably be traced back to Hopfield, a Chicago-born emeritus professor at Princeton University.
Physicist, molecular biologist and neuroscientist John J. Hopfield.
In 1982, he invented the “Hopfield network,” a type of neural network — as these machine-learning programs are known — that was capable of mirroring certain functions of the human brain and recalling “memories” using only partial information.
Hinton is a British-Canadian professor at the University of Toronto who is often referred to as one of the “godfathers of AI.” He used Hopfield’s invention to come up with his own network capable of recognizing shared characteristics among large sets of data. An everyday use for this might be classifying lots of images based on things contained within them.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection,” Hinton said Tuesday, quoted by the Royal Swedish Academy of Sciences that announced the prize. “I was going to have an MRI scan today but I’ll have to cancel that!”
He worked for a decade at Google, becoming one of the world’s most renowned voices on AI. He very publicly quit his job last May, posting on X that he made the decision “so that I could talk about the dangers of AI.”
“It is hard to see how you can prevent the bad actors from using it for bad things,” Hinton said in an interview with The New York Times. | 5 |
The Royal Swedish Academy of Sciences announced the prize in Stockholm.
An American professor and a British-Canadian professor won the Nobel Prize in Physics on Tuesday for their decadeslong, trailblazing research forming the building blocks of artificial intelligence.
John J. Hopfield, 91, was awarded the honor alongside Geoffrey E. Hinton, 76, who left his job at Google last year so he could speak freely about his concerns over the technology.
The pair are central figures in the creation of modern-day AI.
Since the 1980s, they have been using tools from physics to develop the foundations of what is known as “machine learning,” in which computers are fed masses of data to learn an array of tasks — from diagnosing diseases to knowing people’s favorite streaming shows.
Their research “formed the building blocks of machine learning, that can aid humans in making faster and more reliable decisions,” Ellen Moons, chair of the Nobel Committee for Physics, told a news conference.
The use of this technology has “become part of our daily lives, for instance in facial recognition and language translation,” Moons said, while warning that AI’s “rapid development has also raised concerns about our future.”
The machine-learning revolution can arguably be traced back to Hopfield, a Chicago-born emeritus professor at Princeton University.
Physicist, molecular biologist and neuroscientist John J. Hopfield.
In 1982, he invented the “Hopfield network,” a type of neural network — as these machine-learning programs are known — that was capable of mirroring certain functions of the human brain and recalling “memories” using only partial information.
Hinton is a British-Canadian professor at the University of Toronto who is often referred to as one of the “godfathers of AI.” He used Hopfield’s invention to come up with his own network capable of recognizing shared characteristics among large sets of data. An everyday use for this might be classifying lots of images based on things contained within them.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection,” Hinton said Tuesday, quoted by the Royal Swedish Academy of Sciences that announced the prize. “I was going to have an MRI scan today but I’ll have to cancel that!”
He worked for a decade at Google, becoming one of the world’s most renowned voices on AI. He very publicly quit his job last May, posting on X that he made the decision “so that I could talk about the dangers of AI.”
“It is hard to see how you can prevent the bad actors from using it for bad things,” Hinton said in an interview with The New York Times. | 5 |
AI research in the UK is vulnerable to nation-state hackers seeking to steal sensitive data and insights, a new report by the Alan Turing Institute has warned.
The Institute urged the government and academia to develop a long-term strategy to address systemic cultural and structural security barriers to effective AI research security.
The researchers noted that the UK’s “world-leading AI research ecosystem” is a high priority target for state threat actors who are looking to use the technology for malicious purposes.
Access to the underlying sensitive datasets used to train AI model could also provide strategic insights that impact defense planning and intelligence efforts.
China, Russia, North Korea and Iran are the states that pose the great threat to AI academic research, the Institute said.
Barriers to AI Research Security
Despite these risks, there are currently major constraints to AI research cybersecurity. These are creating opportunities for state threat actors to acquire knowledge or steal intellectual property (IP).
Much of this is born out of a “fundamental tension” between academic freedom and research security, according to the researchers.
Cultural Resistance in Academia
The researchers noted that academics are under significant pressure to provide transparency about the data and methods they used to make their findings. Academic journals will often reject submissions where data and code are not made available.
These transparency practices “embed an inherent vulnerability in academic culture,” as threat actors can use this underlying data and described techniques for malicious purposes.
Informal peer-to-peer academic collaborations compound this issue due to the information-sharing culture of academia in early-stage research, the researchers added.
Existing Procedures Are Restrictive
The report also found that academic research security can be more resource-intensive than other forms of due diligence because of the myriad considerations required to understand the potential risks.
This includes the large number of government departments involved in research security, which creates friction for academics and professional services staff seeking guidance.
This friction has resulted in a lack of incentives for researchers to follow non-binding government-issued security guidance.
Lack of Security Awareness
Another major barrier is the lack awareness of the security threat to AI research within the academic community.
Individual academics often have to make personal judgements on the risks of their research, which can be challenging to do in practice.
“It is difficult for researchers to foresee and quantify the risks stemming from early-stage research – and understanding how research may be exploited by adversaries is not an easy task,” the report noted.
Academia’s Funding and Talent Shortage
A lack of access to funding and poor talent retention in academia also introduces new vulnerabilities relating to research security.
Academics are sometimes incentivized to accept funding from dubious sources or accept higher-paid roles at organizations that that can then exploit their insight and expertise.
These organizations may be linked to nation-states with malicious intentions around AI research.
Striking the Balance Between Security and Academic Freedom
The report provided several recommendations for the UK government and academia to strike a balance between the open nature of academic AI research and effective research security practices.
These include: | 5 |
AI research in the UK is vulnerable to nation-state hackers seeking to steal sensitive data and insights, a new report by the Alan Turing Institute has warned.
The Institute urged the government and academia to develop a long-term strategy to address systemic cultural and structural security barriers to effective AI research security.
The researchers noted that the UK’s “world-leading AI research ecosystem” is a high priority target for state threat actors who are looking to use the technology for malicious purposes.
Access to the underlying sensitive datasets used to train AI model could also provide strategic insights that impact defense planning and intelligence efforts.
China, Russia, North Korea and Iran are the states that pose the great threat to AI academic research, the Institute said.
Barriers to AI Research Security
Despite these risks, there are currently major constraints to AI research cybersecurity. These are creating opportunities for state threat actors to acquire knowledge or steal intellectual property (IP).
Much of this is born out of a “fundamental tension” between academic freedom and research security, according to the researchers.
Cultural Resistance in Academia
The researchers noted that academics are under significant pressure to provide transparency about the data and methods they used to make their findings. Academic journals will often reject submissions where data and code are not made available.
These transparency practices “embed an inherent vulnerability in academic culture,” as threat actors can use this underlying data and described techniques for malicious purposes.
Informal peer-to-peer academic collaborations compound this issue due to the information-sharing culture of academia in early-stage research, the researchers added.
Existing Procedures Are Restrictive
The report also found that academic research security can be more resource-intensive than other forms of due diligence because of the myriad considerations required to understand the potential risks.
This includes the large number of government departments involved in research security, which creates friction for academics and professional services staff seeking guidance.
This friction has resulted in a lack of incentives for researchers to follow non-binding government-issued security guidance.
Lack of Security Awareness
Another major barrier is the lack awareness of the security threat to AI research within the academic community.
Individual academics often have to make personal judgements on the risks of their research, which can be challenging to do in practice.
“It is difficult for researchers to foresee and quantify the risks stemming from early-stage research – and understanding how research may be exploited by adversaries is not an easy task,” the report noted.
Academia’s Funding and Talent Shortage
A lack of access to funding and poor talent retention in academia also introduces new vulnerabilities relating to research security.
Academics are sometimes incentivized to accept funding from dubious sources or accept higher-paid roles at organizations that that can then exploit their insight and expertise.
These organizations may be linked to nation-states with malicious intentions around AI research.
Striking the Balance Between Security and Academic Freedom
The report provided several recommendations for the UK government and academia to strike a balance between the open nature of academic AI research and effective research security practices.
These include: | 5 |
AI research in the UK is vulnerable to nation-state hackers seeking to steal sensitive data and insights, a new report by the Alan Turing Institute has warned.
The Institute urged the government and academia to develop a long-term strategy to address systemic cultural and structural security barriers to effective AI research security.
The researchers noted that the UK’s “world-leading AI research ecosystem” is a high priority target for state threat actors who are looking to use the technology for malicious purposes.
Access to the underlying sensitive datasets used to train AI model could also provide strategic insights that impact defense planning and intelligence efforts.
China, Russia, North Korea and Iran are the states that pose the great threat to AI academic research, the Institute said.
Barriers to AI Research Security
Despite these risks, there are currently major constraints to AI research cybersecurity. These are creating opportunities for state threat actors to acquire knowledge or steal intellectual property (IP).
Much of this is born out of a “fundamental tension” between academic freedom and research security, according to the researchers.
Cultural Resistance in Academia
The researchers noted that academics are under significant pressure to provide transparency about the data and methods they used to make their findings. Academic journals will often reject submissions where data and code are not made available.
These transparency practices “embed an inherent vulnerability in academic culture,” as threat actors can use this underlying data and described techniques for malicious purposes.
Informal peer-to-peer academic collaborations compound this issue due to the information-sharing culture of academia in early-stage research, the researchers added.
Existing Procedures Are Restrictive
The report also found that academic research security can be more resource-intensive than other forms of due diligence because of the myriad considerations required to understand the potential risks.
This includes the large number of government departments involved in research security, which creates friction for academics and professional services staff seeking guidance.
This friction has resulted in a lack of incentives for researchers to follow non-binding government-issued security guidance.
Lack of Security Awareness
Another major barrier is the lack awareness of the security threat to AI research within the academic community.
Individual academics often have to make personal judgements on the risks of their research, which can be challenging to do in practice.
“It is difficult for researchers to foresee and quantify the risks stemming from early-stage research – and understanding how research may be exploited by adversaries is not an easy task,” the report noted.
Academia’s Funding and Talent Shortage
A lack of access to funding and poor talent retention in academia also introduces new vulnerabilities relating to research security.
Academics are sometimes incentivized to accept funding from dubious sources or accept higher-paid roles at organizations that that can then exploit their insight and expertise.
These organizations may be linked to nation-states with malicious intentions around AI research.
Striking the Balance Between Security and Academic Freedom
The report provided several recommendations for the UK government and academia to strike a balance between the open nature of academic AI research and effective research security practices.
These include: | 5 |
AI research in the UK is vulnerable to nation-state hackers seeking to steal sensitive data and insights, a new report by the Alan Turing Institute has warned.
The Institute urged the government and academia to develop a long-term strategy to address systemic cultural and structural security barriers to effective AI research security.
The researchers noted that the UK’s “world-leading AI research ecosystem” is a high priority target for state threat actors who are looking to use the technology for malicious purposes.
Access to the underlying sensitive datasets used to train AI model could also provide strategic insights that impact defense planning and intelligence efforts.
China, Russia, North Korea and Iran are the states that pose the great threat to AI academic research, the Institute said.
Barriers to AI Research Security
Despite these risks, there are currently major constraints to AI research cybersecurity. These are creating opportunities for state threat actors to acquire knowledge or steal intellectual property (IP).
Much of this is born out of a “fundamental tension” between academic freedom and research security, according to the researchers.
Cultural Resistance in Academia
The researchers noted that academics are under significant pressure to provide transparency about the data and methods they used to make their findings. Academic journals will often reject submissions where data and code are not made available.
These transparency practices “embed an inherent vulnerability in academic culture,” as threat actors can use this underlying data and described techniques for malicious purposes.
Informal peer-to-peer academic collaborations compound this issue due to the information-sharing culture of academia in early-stage research, the researchers added.
Existing Procedures Are Restrictive
The report also found that academic research security can be more resource-intensive than other forms of due diligence because of the myriad considerations required to understand the potential risks.
This includes the large number of government departments involved in research security, which creates friction for academics and professional services staff seeking guidance.
This friction has resulted in a lack of incentives for researchers to follow non-binding government-issued security guidance.
Lack of Security Awareness
Another major barrier is the lack awareness of the security threat to AI research within the academic community.
Individual academics often have to make personal judgements on the risks of their research, which can be challenging to do in practice.
“It is difficult for researchers to foresee and quantify the risks stemming from early-stage research – and understanding how research may be exploited by adversaries is not an easy task,” the report noted.
Academia’s Funding and Talent Shortage
A lack of access to funding and poor talent retention in academia also introduces new vulnerabilities relating to research security.
Academics are sometimes incentivized to accept funding from dubious sources or accept higher-paid roles at organizations that that can then exploit their insight and expertise.
These organizations may be linked to nation-states with malicious intentions around AI research.
Striking the Balance Between Security and Academic Freedom
The report provided several recommendations for the UK government and academia to strike a balance between the open nature of academic AI research and effective research security practices.
These include: | 5 |
AI research in the UK is vulnerable to nation-state hackers seeking to steal sensitive data and insights, a new report by the Alan Turing Institute has warned.
The Institute urged the government and academia to develop a long-term strategy to address systemic cultural and structural security barriers to effective AI research security.
The researchers noted that the UK’s “world-leading AI research ecosystem” is a high priority target for state threat actors who are looking to use the technology for malicious purposes.
Access to the underlying sensitive datasets used to train AI model could also provide strategic insights that impact defense planning and intelligence efforts.
China, Russia, North Korea and Iran are the states that pose the great threat to AI academic research, the Institute said.
Barriers to AI Research Security
Despite these risks, there are currently major constraints to AI research cybersecurity. These are creating opportunities for state threat actors to acquire knowledge or steal intellectual property (IP).
Much of this is born out of a “fundamental tension” between academic freedom and research security, according to the researchers.
Cultural Resistance in Academia
The researchers noted that academics are under significant pressure to provide transparency about the data and methods they used to make their findings. Academic journals will often reject submissions where data and code are not made available.
These transparency practices “embed an inherent vulnerability in academic culture,” as threat actors can use this underlying data and described techniques for malicious purposes.
Informal peer-to-peer academic collaborations compound this issue due to the information-sharing culture of academia in early-stage research, the researchers added.
Existing Procedures Are Restrictive
The report also found that academic research security can be more resource-intensive than other forms of due diligence because of the myriad considerations required to understand the potential risks.
This includes the large number of government departments involved in research security, which creates friction for academics and professional services staff seeking guidance.
This friction has resulted in a lack of incentives for researchers to follow non-binding government-issued security guidance.
Lack of Security Awareness
Another major barrier is the lack awareness of the security threat to AI research within the academic community.
Individual academics often have to make personal judgements on the risks of their research, which can be challenging to do in practice.
“It is difficult for researchers to foresee and quantify the risks stemming from early-stage research – and understanding how research may be exploited by adversaries is not an easy task,” the report noted.
Academia’s Funding and Talent Shortage
A lack of access to funding and poor talent retention in academia also introduces new vulnerabilities relating to research security.
Academics are sometimes incentivized to accept funding from dubious sources or accept higher-paid roles at organizations that that can then exploit their insight and expertise.
These organizations may be linked to nation-states with malicious intentions around AI research.
Striking the Balance Between Security and Academic Freedom
The report provided several recommendations for the UK government and academia to strike a balance between the open nature of academic AI research and effective research security practices.
These include: | 5 |
New Northeastern research tests the capabilities of AI chatbots on NPR’s Sunday Puzzles New Northeastern research reveals how capable of AI chatbots are at solving NPR’s Sunday Puzzle. Photo illustration by Matthew Modoono/Northeastern University
Listeners of NPR’s Sunday Puzzle are well aware just how challenging the weekly quiz show can be, requiring participants to have a strong grasp of popular culture and the English language.
While the puzzles may not be the easiest to solve, they aren’t impossible. With some thinking and trial and error, everyday people answer them correctly every week.
That’s what made them the perfect data source for a new benchmark researchers have developed to test the capabilities of the latest artificial intelligence reasoning models coming out of OpenAI, Google, Anthropic and DeepSeek.
Arjun Guha, associate professor of computer science at Northeastern University, is one of the co-author of the benchmark study. Photo by Matthew Modoono/Northeastern University
It’s a common practice for AI researchers working in the field to develop specific benchmarks to measure progress and the capabilities of AI technologies, explains Arjun Guha, a Northeastern University professor in the Khoury College of Computer Sciences and one of the authors of the research.
The issue, however, is that the models have become so advanced that the tasks they are asked have become more challenging to accomplish and measure.
“You have questions that are very narrowly designed by Ph.D. students and are only answerable by people with Ph.D.s in a narrow field of expertise,” he says.
The questions asked during NPR’s Sunday Puzzles, on the other hand, while difficult, can be easily identifiable by nonexperts.
“You can really look at them as a test of verbal reasoning skills and general knowledge,” says Guha. “There’s a lot of ‘find a five-letter word with the following letter-by-letter properties, and it’s the name of some obscure city or some movie from the ’80s or something.”
For the study, researchers tested out a new crop of reasoning models released by OpenAI, Google, Anthropic and DeepSeek in the past few months. What sets reasoning models apart is that they are trained with reinforcement learning techniques and “show their work,” meaning they explain step by step how they come up with their answers. | 5 |
New Northeastern research tests the capabilities of AI chatbots on NPR’s Sunday Puzzles New Northeastern research reveals how capable of AI chatbots are at solving NPR’s Sunday Puzzle. Photo illustration by Matthew Modoono/Northeastern University
Listeners of NPR’s Sunday Puzzle are well aware just how challenging the weekly quiz show can be, requiring participants to have a strong grasp of popular culture and the English language.
While the puzzles may not be the easiest to solve, they aren’t impossible. With some thinking and trial and error, everyday people answer them correctly every week.
That’s what made them the perfect data source for a new benchmark researchers have developed to test the capabilities of the latest artificial intelligence reasoning models coming out of OpenAI, Google, Anthropic and DeepSeek.
Arjun Guha, associate professor of computer science at Northeastern University, is one of the co-author of the benchmark study. Photo by Matthew Modoono/Northeastern University
It’s a common practice for AI researchers working in the field to develop specific benchmarks to measure progress and the capabilities of AI technologies, explains Arjun Guha, a Northeastern University professor in the Khoury College of Computer Sciences and one of the authors of the research.
The issue, however, is that the models have become so advanced that the tasks they are asked have become more challenging to accomplish and measure.
“You have questions that are very narrowly designed by Ph.D. students and are only answerable by people with Ph.D.s in a narrow field of expertise,” he says.
The questions asked during NPR’s Sunday Puzzles, on the other hand, while difficult, can be easily identifiable by nonexperts.
“You can really look at them as a test of verbal reasoning skills and general knowledge,” says Guha. “There’s a lot of ‘find a five-letter word with the following letter-by-letter properties, and it’s the name of some obscure city or some movie from the ’80s or something.”
For the study, researchers tested out a new crop of reasoning models released by OpenAI, Google, Anthropic and DeepSeek in the past few months. What sets reasoning models apart is that they are trained with reinforcement learning techniques and “show their work,” meaning they explain step by step how they come up with their answers. | 5 |
New Northeastern research tests the capabilities of AI chatbots on NPR’s Sunday Puzzles New Northeastern research reveals how capable of AI chatbots are at solving NPR’s Sunday Puzzle. Photo illustration by Matthew Modoono/Northeastern University
Listeners of NPR’s Sunday Puzzle are well aware just how challenging the weekly quiz show can be, requiring participants to have a strong grasp of popular culture and the English language.
While the puzzles may not be the easiest to solve, they aren’t impossible. With some thinking and trial and error, everyday people answer them correctly every week.
That’s what made them the perfect data source for a new benchmark researchers have developed to test the capabilities of the latest artificial intelligence reasoning models coming out of OpenAI, Google, Anthropic and DeepSeek.
Arjun Guha, associate professor of computer science at Northeastern University, is one of the co-author of the benchmark study. Photo by Matthew Modoono/Northeastern University
It’s a common practice for AI researchers working in the field to develop specific benchmarks to measure progress and the capabilities of AI technologies, explains Arjun Guha, a Northeastern University professor in the Khoury College of Computer Sciences and one of the authors of the research.
The issue, however, is that the models have become so advanced that the tasks they are asked have become more challenging to accomplish and measure.
“You have questions that are very narrowly designed by Ph.D. students and are only answerable by people with Ph.D.s in a narrow field of expertise,” he says.
The questions asked during NPR’s Sunday Puzzles, on the other hand, while difficult, can be easily identifiable by nonexperts.
“You can really look at them as a test of verbal reasoning skills and general knowledge,” says Guha. “There’s a lot of ‘find a five-letter word with the following letter-by-letter properties, and it’s the name of some obscure city or some movie from the ’80s or something.”
For the study, researchers tested out a new crop of reasoning models released by OpenAI, Google, Anthropic and DeepSeek in the past few months. What sets reasoning models apart is that they are trained with reinforcement learning techniques and “show their work,” meaning they explain step by step how they come up with their answers. | 5 |
New Northeastern research tests the capabilities of AI chatbots on NPR’s Sunday Puzzles New Northeastern research reveals how capable of AI chatbots are at solving NPR’s Sunday Puzzle. Photo illustration by Matthew Modoono/Northeastern University
Listeners of NPR’s Sunday Puzzle are well aware just how challenging the weekly quiz show can be, requiring participants to have a strong grasp of popular culture and the English language.
While the puzzles may not be the easiest to solve, they aren’t impossible. With some thinking and trial and error, everyday people answer them correctly every week.
That’s what made them the perfect data source for a new benchmark researchers have developed to test the capabilities of the latest artificial intelligence reasoning models coming out of OpenAI, Google, Anthropic and DeepSeek.
Arjun Guha, associate professor of computer science at Northeastern University, is one of the co-author of the benchmark study. Photo by Matthew Modoono/Northeastern University
It’s a common practice for AI researchers working in the field to develop specific benchmarks to measure progress and the capabilities of AI technologies, explains Arjun Guha, a Northeastern University professor in the Khoury College of Computer Sciences and one of the authors of the research.
The issue, however, is that the models have become so advanced that the tasks they are asked have become more challenging to accomplish and measure.
“You have questions that are very narrowly designed by Ph.D. students and are only answerable by people with Ph.D.s in a narrow field of expertise,” he says.
The questions asked during NPR’s Sunday Puzzles, on the other hand, while difficult, can be easily identifiable by nonexperts.
“You can really look at them as a test of verbal reasoning skills and general knowledge,” says Guha. “There’s a lot of ‘find a five-letter word with the following letter-by-letter properties, and it’s the name of some obscure city or some movie from the ’80s or something.”
For the study, researchers tested out a new crop of reasoning models released by OpenAI, Google, Anthropic and DeepSeek in the past few months. What sets reasoning models apart is that they are trained with reinforcement learning techniques and “show their work,” meaning they explain step by step how they come up with their answers. | 5 |
New Northeastern research tests the capabilities of AI chatbots on NPR’s Sunday Puzzles New Northeastern research reveals how capable of AI chatbots are at solving NPR’s Sunday Puzzle. Photo illustration by Matthew Modoono/Northeastern University
Listeners of NPR’s Sunday Puzzle are well aware just how challenging the weekly quiz show can be, requiring participants to have a strong grasp of popular culture and the English language.
While the puzzles may not be the easiest to solve, they aren’t impossible. With some thinking and trial and error, everyday people answer them correctly every week.
That’s what made them the perfect data source for a new benchmark researchers have developed to test the capabilities of the latest artificial intelligence reasoning models coming out of OpenAI, Google, Anthropic and DeepSeek.
Arjun Guha, associate professor of computer science at Northeastern University, is one of the co-author of the benchmark study. Photo by Matthew Modoono/Northeastern University
It’s a common practice for AI researchers working in the field to develop specific benchmarks to measure progress and the capabilities of AI technologies, explains Arjun Guha, a Northeastern University professor in the Khoury College of Computer Sciences and one of the authors of the research.
The issue, however, is that the models have become so advanced that the tasks they are asked have become more challenging to accomplish and measure.
“You have questions that are very narrowly designed by Ph.D. students and are only answerable by people with Ph.D.s in a narrow field of expertise,” he says.
The questions asked during NPR’s Sunday Puzzles, on the other hand, while difficult, can be easily identifiable by nonexperts.
“You can really look at them as a test of verbal reasoning skills and general knowledge,” says Guha. “There’s a lot of ‘find a five-letter word with the following letter-by-letter properties, and it’s the name of some obscure city or some movie from the ’80s or something.”
For the study, researchers tested out a new crop of reasoning models released by OpenAI, Google, Anthropic and DeepSeek in the past few months. What sets reasoning models apart is that they are trained with reinforcement learning techniques and “show their work,” meaning they explain step by step how they come up with their answers. | 5 |
The University of Mississippi announced last week that the university is among a small group of institutions across the globe invited to join NextGenAI – a $50 million initiative that aims to advance artificial intelligence research and education.
The initiative is headed by OpenAI, who announced the consortium on March 4. Ole Miss is the only university in Mississippi to be selected for the collaboration and one of three Southeastern Conference schools.
NextGenAI will provide a $50,000 grant to Ole Miss to fund research through the National Center of Narrative Intelligence in partnership with the Center for Practical Ethics and the Institute for Data Science.
“This grant is going to allow us to support interdisciplinary, collaborative research,” Wes Jennings, co-director of the National Center for Narrative Intelligence (NCNI) at Ole Miss. “We want to kick-start ideas and provide resources that can help those working with AI with campus.”
The collaboration, which includes Duke University, Ohio State University, and Harvard University, will connect 15 total institutions to support AI-driven research in multiple fields. The research will focus primarily on driving progress in science, medicine, technology, and education.
“A close collaboration with universities is essential to our mission of building AI that benefits everyone,” Brad Lightcap, chief operating officer of OpenAI, said. “NextGenAI will accelerate research progress and catalyze a new generation of institutions equipped to harness to the transformative power of AI.”
The NCNI will hold a competitive grant process to identify projects by Ole Miss researchers that would benefit from OpenAI’s support. Jennings believes the research can make a big difference in a changing world.
“We know that AI can do things in minutes that would take researchers hundreds of hours to do by hand,” Jennings said. “It’s a tool; let’s use it to solve grand challenges nimbly and with efficiency. We’re looking for bright ideas and bright minds to work on them.” | 5 |
The University of Mississippi announced last week that the university is among a small group of institutions across the globe invited to join NextGenAI – a $50 million initiative that aims to advance artificial intelligence research and education.
The initiative is headed by OpenAI, who announced the consortium on March 4. Ole Miss is the only university in Mississippi to be selected for the collaboration and one of three Southeastern Conference schools.
NextGenAI will provide a $50,000 grant to Ole Miss to fund research through the National Center of Narrative Intelligence in partnership with the Center for Practical Ethics and the Institute for Data Science.
“This grant is going to allow us to support interdisciplinary, collaborative research,” Wes Jennings, co-director of the National Center for Narrative Intelligence (NCNI) at Ole Miss. “We want to kick-start ideas and provide resources that can help those working with AI with campus.”
The collaboration, which includes Duke University, Ohio State University, and Harvard University, will connect 15 total institutions to support AI-driven research in multiple fields. The research will focus primarily on driving progress in science, medicine, technology, and education.
“A close collaboration with universities is essential to our mission of building AI that benefits everyone,” Brad Lightcap, chief operating officer of OpenAI, said. “NextGenAI will accelerate research progress and catalyze a new generation of institutions equipped to harness to the transformative power of AI.”
The NCNI will hold a competitive grant process to identify projects by Ole Miss researchers that would benefit from OpenAI’s support. Jennings believes the research can make a big difference in a changing world.
“We know that AI can do things in minutes that would take researchers hundreds of hours to do by hand,” Jennings said. “It’s a tool; let’s use it to solve grand challenges nimbly and with efficiency. We’re looking for bright ideas and bright minds to work on them.” | 5 |
The University of Mississippi announced last week that the university is among a small group of institutions across the globe invited to join NextGenAI – a $50 million initiative that aims to advance artificial intelligence research and education.
The initiative is headed by OpenAI, who announced the consortium on March 4. Ole Miss is the only university in Mississippi to be selected for the collaboration and one of three Southeastern Conference schools.
NextGenAI will provide a $50,000 grant to Ole Miss to fund research through the National Center of Narrative Intelligence in partnership with the Center for Practical Ethics and the Institute for Data Science.
“This grant is going to allow us to support interdisciplinary, collaborative research,” Wes Jennings, co-director of the National Center for Narrative Intelligence (NCNI) at Ole Miss. “We want to kick-start ideas and provide resources that can help those working with AI with campus.”
The collaboration, which includes Duke University, Ohio State University, and Harvard University, will connect 15 total institutions to support AI-driven research in multiple fields. The research will focus primarily on driving progress in science, medicine, technology, and education.
“A close collaboration with universities is essential to our mission of building AI that benefits everyone,” Brad Lightcap, chief operating officer of OpenAI, said. “NextGenAI will accelerate research progress and catalyze a new generation of institutions equipped to harness to the transformative power of AI.”
The NCNI will hold a competitive grant process to identify projects by Ole Miss researchers that would benefit from OpenAI’s support. Jennings believes the research can make a big difference in a changing world.
“We know that AI can do things in minutes that would take researchers hundreds of hours to do by hand,” Jennings said. “It’s a tool; let’s use it to solve grand challenges nimbly and with efficiency. We’re looking for bright ideas and bright minds to work on them.” | 5 |
The University of Mississippi announced last week that the university is among a small group of institutions across the globe invited to join NextGenAI – a $50 million initiative that aims to advance artificial intelligence research and education.
The initiative is headed by OpenAI, who announced the consortium on March 4. Ole Miss is the only university in Mississippi to be selected for the collaboration and one of three Southeastern Conference schools.
NextGenAI will provide a $50,000 grant to Ole Miss to fund research through the National Center of Narrative Intelligence in partnership with the Center for Practical Ethics and the Institute for Data Science.
“This grant is going to allow us to support interdisciplinary, collaborative research,” Wes Jennings, co-director of the National Center for Narrative Intelligence (NCNI) at Ole Miss. “We want to kick-start ideas and provide resources that can help those working with AI with campus.”
The collaboration, which includes Duke University, Ohio State University, and Harvard University, will connect 15 total institutions to support AI-driven research in multiple fields. The research will focus primarily on driving progress in science, medicine, technology, and education.
“A close collaboration with universities is essential to our mission of building AI that benefits everyone,” Brad Lightcap, chief operating officer of OpenAI, said. “NextGenAI will accelerate research progress and catalyze a new generation of institutions equipped to harness to the transformative power of AI.”
The NCNI will hold a competitive grant process to identify projects by Ole Miss researchers that would benefit from OpenAI’s support. Jennings believes the research can make a big difference in a changing world.
“We know that AI can do things in minutes that would take researchers hundreds of hours to do by hand,” Jennings said. “It’s a tool; let’s use it to solve grand challenges nimbly and with efficiency. We’re looking for bright ideas and bright minds to work on them.” | 5 |
The University of Mississippi announced last week that the university is among a small group of institutions across the globe invited to join NextGenAI – a $50 million initiative that aims to advance artificial intelligence research and education.
The initiative is headed by OpenAI, who announced the consortium on March 4. Ole Miss is the only university in Mississippi to be selected for the collaboration and one of three Southeastern Conference schools.
NextGenAI will provide a $50,000 grant to Ole Miss to fund research through the National Center of Narrative Intelligence in partnership with the Center for Practical Ethics and the Institute for Data Science.
“This grant is going to allow us to support interdisciplinary, collaborative research,” Wes Jennings, co-director of the National Center for Narrative Intelligence (NCNI) at Ole Miss. “We want to kick-start ideas and provide resources that can help those working with AI with campus.”
The collaboration, which includes Duke University, Ohio State University, and Harvard University, will connect 15 total institutions to support AI-driven research in multiple fields. The research will focus primarily on driving progress in science, medicine, technology, and education.
“A close collaboration with universities is essential to our mission of building AI that benefits everyone,” Brad Lightcap, chief operating officer of OpenAI, said. “NextGenAI will accelerate research progress and catalyze a new generation of institutions equipped to harness to the transformative power of AI.”
The NCNI will hold a competitive grant process to identify projects by Ole Miss researchers that would benefit from OpenAI’s support. Jennings believes the research can make a big difference in a changing world.
“We know that AI can do things in minutes that would take researchers hundreds of hours to do by hand,” Jennings said. “It’s a tool; let’s use it to solve grand challenges nimbly and with efficiency. We’re looking for bright ideas and bright minds to work on them.” | 5 |
SRM University-AP (SRM-AP), Andhra Pradesh, is proud to announce a transformative five-year collaboration with Carnegie Mellon University’s School of Computer Science (CMU SCS), USA- one of the world’s foremost institutions in artificial intelligence (AI) and cutting-edge research. This strategic collaboration aims to push the boundaries of knowledge, innovation and education in AI- related disciplines, including machine learning, natural language processing, computer vision, infrastructure and systems, and AI ethics and policy.
At the heart of this collaboration is a shared vision to foster an ecosystem that nurtures groundbreaking research, cultivates exceptional talent and accelerates advancements in AI-driven technologies.
A Pioneering Collaboration for AI Excellence
“CMU’s School of Computer Science is excited to work with SRM University-AP on this landmark collaboration to advance research and bolster AI education. Together, we will shape the future of AI and empower the next generation of researchers, educators and industry leaders to push the frontiers of technology and drive meaningful change in society,” said Prof. Martial Hebert, Dean of CMU's School of Computer Science.
Empowering Research Through Global Collaboration
As part of this collaboration, SRM University-AP’s research faculty and researchers will have the opportunity to engage directly with the esteemed faculty and researchers at CMU’s School of Computer Science. They will immerse themselves in CMU SCS’s pioneering AI labs, working alongside global experts in key research domains. This will facilitate research, knowledge sharing and the development of state-of-the-art AI innovations that address real-world challenges.
Dr P Sathyanarayanan, Pro-Chancellor of SRM University-AP said that “To further strengthen research capabilities, this collaboration will also pave the way to establish advanced AI labs at SRM University-AP. These labs will be incubators for novel AI research, fostering a stimulating environment that promotes academic rigor, interdisciplinary collaboration and technological innovation”.
Advancing AI Education with World-Class Learning Opportunities
Beyond research, this collaboration is designed to enrich the academic experience of SRM-AP’s teaching faculty and research scholars. Selected faculty members and scholars can audit cutting-edge AI courses at CMU’s School of Computer Science as visiting participants. This exposure will allow them to engage with CMU SCS faculty and contribute to developing robust AI curricula at SRM-AP. They will also gain hands-on experience in designing assignments, worksheets and examinations that mirror real-world AI problem-solving scenarios, enhancing the quality of AI education at SRM University-AP.
Unparalleled Research Internships for Students
Prof. Manoj K Arora, Vice Chancellor of SRM University-AP expressed that “In a move that underscores its commitment to nurturing future AI leaders, the collaboration will offer SRM-AP students the opportunity to undertake research internships at CMU’s School of Computer Science.” Selected students will spend approx. six weeks each summer immersed in a world-class research environment, gaining firsthand experience in tackling complex AI challenges alongside leaders in the field. This experience will provide students with unparalleled insights and exposure to global research methodologies, setting them apart in the highly competitive AI landscape.
By leveraging CMU SCS’s expertise and SRM-AP’s commitment to academic excellence, this collaboration will drive innovation, expand knowledge horizons and create a lasting impact on the AI ecosystem between the universities. | 5 |
SRM University-AP (SRM-AP), Andhra Pradesh, is proud to announce a transformative five-year collaboration with Carnegie Mellon University’s School of Computer Science (CMU SCS), USA- one of the world’s foremost institutions in artificial intelligence (AI) and cutting-edge research. This strategic collaboration aims to push the boundaries of knowledge, innovation and education in AI- related disciplines, including machine learning, natural language processing, computer vision, infrastructure and systems, and AI ethics and policy.
At the heart of this collaboration is a shared vision to foster an ecosystem that nurtures groundbreaking research, cultivates exceptional talent and accelerates advancements in AI-driven technologies.
A Pioneering Collaboration for AI Excellence
“CMU’s School of Computer Science is excited to work with SRM University-AP on this landmark collaboration to advance research and bolster AI education. Together, we will shape the future of AI and empower the next generation of researchers, educators and industry leaders to push the frontiers of technology and drive meaningful change in society,” said Prof. Martial Hebert, Dean of CMU's School of Computer Science.
Empowering Research Through Global Collaboration
As part of this collaboration, SRM University-AP’s research faculty and researchers will have the opportunity to engage directly with the esteemed faculty and researchers at CMU’s School of Computer Science. They will immerse themselves in CMU SCS’s pioneering AI labs, working alongside global experts in key research domains. This will facilitate research, knowledge sharing and the development of state-of-the-art AI innovations that address real-world challenges.
Dr P Sathyanarayanan, Pro-Chancellor of SRM University-AP said that “To further strengthen research capabilities, this collaboration will also pave the way to establish advanced AI labs at SRM University-AP. These labs will be incubators for novel AI research, fostering a stimulating environment that promotes academic rigor, interdisciplinary collaboration and technological innovation”.
Advancing AI Education with World-Class Learning Opportunities
Beyond research, this collaboration is designed to enrich the academic experience of SRM-AP’s teaching faculty and research scholars. Selected faculty members and scholars can audit cutting-edge AI courses at CMU’s School of Computer Science as visiting participants. This exposure will allow them to engage with CMU SCS faculty and contribute to developing robust AI curricula at SRM-AP. They will also gain hands-on experience in designing assignments, worksheets and examinations that mirror real-world AI problem-solving scenarios, enhancing the quality of AI education at SRM University-AP.
Unparalleled Research Internships for Students
Prof. Manoj K Arora, Vice Chancellor of SRM University-AP expressed that “In a move that underscores its commitment to nurturing future AI leaders, the collaboration will offer SRM-AP students the opportunity to undertake research internships at CMU’s School of Computer Science.” Selected students will spend approx. six weeks each summer immersed in a world-class research environment, gaining firsthand experience in tackling complex AI challenges alongside leaders in the field. This experience will provide students with unparalleled insights and exposure to global research methodologies, setting them apart in the highly competitive AI landscape.
By leveraging CMU SCS’s expertise and SRM-AP’s commitment to academic excellence, this collaboration will drive innovation, expand knowledge horizons and create a lasting impact on the AI ecosystem between the universities. | 5 |
SRM University-AP (SRM-AP), Andhra Pradesh, is proud to announce a transformative five-year collaboration with Carnegie Mellon University’s School of Computer Science (CMU SCS), USA- one of the world’s foremost institutions in artificial intelligence (AI) and cutting-edge research. This strategic collaboration aims to push the boundaries of knowledge, innovation and education in AI- related disciplines, including machine learning, natural language processing, computer vision, infrastructure and systems, and AI ethics and policy.
At the heart of this collaboration is a shared vision to foster an ecosystem that nurtures groundbreaking research, cultivates exceptional talent and accelerates advancements in AI-driven technologies.
A Pioneering Collaboration for AI Excellence
“CMU’s School of Computer Science is excited to work with SRM University-AP on this landmark collaboration to advance research and bolster AI education. Together, we will shape the future of AI and empower the next generation of researchers, educators and industry leaders to push the frontiers of technology and drive meaningful change in society,” said Prof. Martial Hebert, Dean of CMU's School of Computer Science.
Empowering Research Through Global Collaboration
As part of this collaboration, SRM University-AP’s research faculty and researchers will have the opportunity to engage directly with the esteemed faculty and researchers at CMU’s School of Computer Science. They will immerse themselves in CMU SCS’s pioneering AI labs, working alongside global experts in key research domains. This will facilitate research, knowledge sharing and the development of state-of-the-art AI innovations that address real-world challenges.
Dr P Sathyanarayanan, Pro-Chancellor of SRM University-AP said that “To further strengthen research capabilities, this collaboration will also pave the way to establish advanced AI labs at SRM University-AP. These labs will be incubators for novel AI research, fostering a stimulating environment that promotes academic rigor, interdisciplinary collaboration and technological innovation”.
Advancing AI Education with World-Class Learning Opportunities
Beyond research, this collaboration is designed to enrich the academic experience of SRM-AP’s teaching faculty and research scholars. Selected faculty members and scholars can audit cutting-edge AI courses at CMU’s School of Computer Science as visiting participants. This exposure will allow them to engage with CMU SCS faculty and contribute to developing robust AI curricula at SRM-AP. They will also gain hands-on experience in designing assignments, worksheets and examinations that mirror real-world AI problem-solving scenarios, enhancing the quality of AI education at SRM University-AP.
Unparalleled Research Internships for Students
Prof. Manoj K Arora, Vice Chancellor of SRM University-AP expressed that “In a move that underscores its commitment to nurturing future AI leaders, the collaboration will offer SRM-AP students the opportunity to undertake research internships at CMU’s School of Computer Science.” Selected students will spend approx. six weeks each summer immersed in a world-class research environment, gaining firsthand experience in tackling complex AI challenges alongside leaders in the field. This experience will provide students with unparalleled insights and exposure to global research methodologies, setting them apart in the highly competitive AI landscape.
By leveraging CMU SCS’s expertise and SRM-AP’s commitment to academic excellence, this collaboration will drive innovation, expand knowledge horizons and create a lasting impact on the AI ecosystem between the universities. | 5 |
SRM University-AP (SRM-AP), Andhra Pradesh, is proud to announce a transformative five-year collaboration with Carnegie Mellon University’s School of Computer Science (CMU SCS), USA- one of the world’s foremost institutions in artificial intelligence (AI) and cutting-edge research. This strategic collaboration aims to push the boundaries of knowledge, innovation and education in AI- related disciplines, including machine learning, natural language processing, computer vision, infrastructure and systems, and AI ethics and policy.
At the heart of this collaboration is a shared vision to foster an ecosystem that nurtures groundbreaking research, cultivates exceptional talent and accelerates advancements in AI-driven technologies.
A Pioneering Collaboration for AI Excellence
“CMU’s School of Computer Science is excited to work with SRM University-AP on this landmark collaboration to advance research and bolster AI education. Together, we will shape the future of AI and empower the next generation of researchers, educators and industry leaders to push the frontiers of technology and drive meaningful change in society,” said Prof. Martial Hebert, Dean of CMU's School of Computer Science.
Empowering Research Through Global Collaboration
As part of this collaboration, SRM University-AP’s research faculty and researchers will have the opportunity to engage directly with the esteemed faculty and researchers at CMU’s School of Computer Science. They will immerse themselves in CMU SCS’s pioneering AI labs, working alongside global experts in key research domains. This will facilitate research, knowledge sharing and the development of state-of-the-art AI innovations that address real-world challenges.
Dr P Sathyanarayanan, Pro-Chancellor of SRM University-AP said that “To further strengthen research capabilities, this collaboration will also pave the way to establish advanced AI labs at SRM University-AP. These labs will be incubators for novel AI research, fostering a stimulating environment that promotes academic rigor, interdisciplinary collaboration and technological innovation”.
Advancing AI Education with World-Class Learning Opportunities
Beyond research, this collaboration is designed to enrich the academic experience of SRM-AP’s teaching faculty and research scholars. Selected faculty members and scholars can audit cutting-edge AI courses at CMU’s School of Computer Science as visiting participants. This exposure will allow them to engage with CMU SCS faculty and contribute to developing robust AI curricula at SRM-AP. They will also gain hands-on experience in designing assignments, worksheets and examinations that mirror real-world AI problem-solving scenarios, enhancing the quality of AI education at SRM University-AP.
Unparalleled Research Internships for Students
Prof. Manoj K Arora, Vice Chancellor of SRM University-AP expressed that “In a move that underscores its commitment to nurturing future AI leaders, the collaboration will offer SRM-AP students the opportunity to undertake research internships at CMU’s School of Computer Science.” Selected students will spend approx. six weeks each summer immersed in a world-class research environment, gaining firsthand experience in tackling complex AI challenges alongside leaders in the field. This experience will provide students with unparalleled insights and exposure to global research methodologies, setting them apart in the highly competitive AI landscape.
By leveraging CMU SCS’s expertise and SRM-AP’s commitment to academic excellence, this collaboration will drive innovation, expand knowledge horizons and create a lasting impact on the AI ecosystem between the universities. | 5 |
SRM University-AP (SRM-AP), Andhra Pradesh, is proud to announce a transformative five-year collaboration with Carnegie Mellon University’s School of Computer Science (CMU SCS), USA- one of the world’s foremost institutions in artificial intelligence (AI) and cutting-edge research. This strategic collaboration aims to push the boundaries of knowledge, innovation and education in AI- related disciplines, including machine learning, natural language processing, computer vision, infrastructure and systems, and AI ethics and policy.
At the heart of this collaboration is a shared vision to foster an ecosystem that nurtures groundbreaking research, cultivates exceptional talent and accelerates advancements in AI-driven technologies.
A Pioneering Collaboration for AI Excellence
“CMU’s School of Computer Science is excited to work with SRM University-AP on this landmark collaboration to advance research and bolster AI education. Together, we will shape the future of AI and empower the next generation of researchers, educators and industry leaders to push the frontiers of technology and drive meaningful change in society,” said Prof. Martial Hebert, Dean of CMU's School of Computer Science.
Empowering Research Through Global Collaboration
As part of this collaboration, SRM University-AP’s research faculty and researchers will have the opportunity to engage directly with the esteemed faculty and researchers at CMU’s School of Computer Science. They will immerse themselves in CMU SCS’s pioneering AI labs, working alongside global experts in key research domains. This will facilitate research, knowledge sharing and the development of state-of-the-art AI innovations that address real-world challenges.
Dr P Sathyanarayanan, Pro-Chancellor of SRM University-AP said that “To further strengthen research capabilities, this collaboration will also pave the way to establish advanced AI labs at SRM University-AP. These labs will be incubators for novel AI research, fostering a stimulating environment that promotes academic rigor, interdisciplinary collaboration and technological innovation”.
Advancing AI Education with World-Class Learning Opportunities
Beyond research, this collaboration is designed to enrich the academic experience of SRM-AP’s teaching faculty and research scholars. Selected faculty members and scholars can audit cutting-edge AI courses at CMU’s School of Computer Science as visiting participants. This exposure will allow them to engage with CMU SCS faculty and contribute to developing robust AI curricula at SRM-AP. They will also gain hands-on experience in designing assignments, worksheets and examinations that mirror real-world AI problem-solving scenarios, enhancing the quality of AI education at SRM University-AP.
Unparalleled Research Internships for Students
Prof. Manoj K Arora, Vice Chancellor of SRM University-AP expressed that “In a move that underscores its commitment to nurturing future AI leaders, the collaboration will offer SRM-AP students the opportunity to undertake research internships at CMU’s School of Computer Science.” Selected students will spend approx. six weeks each summer immersed in a world-class research environment, gaining firsthand experience in tackling complex AI challenges alongside leaders in the field. This experience will provide students with unparalleled insights and exposure to global research methodologies, setting them apart in the highly competitive AI landscape.
By leveraging CMU SCS’s expertise and SRM-AP’s commitment to academic excellence, this collaboration will drive innovation, expand knowledge horizons and create a lasting impact on the AI ecosystem between the universities. | 5 |
The field of cancer treatment has long struggled with the immense costs and time-consuming nature of drug development. Traditional methods often take over a decade and billions of dollars to bring a single drug to market, with many compounds failing in late-stage trials due to efficacy or safety concerns. However, artificial intelligence (AI) is now revolutionizing this space by accelerating drug repurposing and designing new therapeutics with unprecedented speed and accuracy. The integration of AI in oncology drug discovery holds the promise of reducing development timelines, optimizing existing drugs, and unveiling novel treatment strategies.
A recent study titled Artificial Intelligence-Based Methods for Drug Repurposing and Development in Cancer, authored by Sara Herráiz-Gil, Elisa Nygren-Jiménez, Diana N. Acosta-Alonso, Carlos León, and Sara Guerrero-Aspizua, and published in Applied Sciences (2025, 15, 2798), presents an in-depth review of AI-driven techniques in drug discovery. The study highlights AI’s role in addressing critical challenges in oncology and explores the latest methodologies and applications in the field.
Role of AI in drug repurposing and new drug development
AI has emerged as a game-changer in the pharmaceutical industry, particularly in oncology, by offering powerful tools for drug repurposing and de novo drug design. Traditional drug repurposing involves identifying new uses for existing drugs, but AI significantly enhances this process by analyzing large-scale biological and chemical data. Machine learning and deep learning algorithms can predict drug-disease interactions, optimize drug efficacy, and minimize toxicity concerns.
The study discusses how knowledge graphs and neural networks are being employed to map complex relationships between drugs and diseases. Graph-based AI techniques allow researchers to identify potential drug candidates based on existing biological networks, while deep learning models can assess molecular interactions with remarkable precision. Generative AI models, such as reinforcement learning algorithms, are also gaining traction in de novo drug design, creating entirely new molecular structures optimized for cancer treatment. By leveraging multi-omics data, AI enables a more precise and personalized approach to therapy.
AI applications in cancer drug discovery
Several groundbreaking AI-driven projects have demonstrated the potential of this technology in oncology. The study outlines multiple case studies where AI was successfully applied to identify promising drug candidates. For instance, AI models have predicted potential therapies for chondrosarcoma, familial melanoma, and liver and lung cancers. By integrating diverse datasets, such as genomic profiles, protein interactions, and clinical trial results, these models provide insights into drug repositioning opportunities that might otherwise remain undiscovered.
Furthermore, AI has accelerated drug screening by predicting the binding affinity of molecules to cancer targets, reducing the need for costly and time-intensive laboratory testing. In hepatocellular carcinoma research, AI-driven de novo drug design has led to the development of a novel CDK20 inhibitor in record time, highlighting the efficiency of computational drug discovery. Additionally, AI-guided strategies have been instrumental in predicting patient responses to specific treatments, paving the way for more targeted and effective cancer therapies.
Experimental integration and challenges in AI-driven drug discovery
While AI presents remarkable opportunities, its integration into traditional experimental workflows comes with challenges. One major limitation is data quality - AI models rely on vast amounts of biomedical data, which can sometimes be inconsistent or biased. Ensuring data standardization and accuracy remains a critical challenge in training reliable AI models.
Another concern is the interpretability of AI predictions. Many deep learning models operate as “black boxes,” making it difficult for researchers to understand the rationale behind certain drug recommendations. To address this, explainable AI (XAI) techniques, such as SHAP and LIME, are being developed to enhance model transparency and regulatory acceptance. Ethical considerations, such as ensuring equitable access to AI-driven therapies and mitigating algorithmic biases, are also central to the responsible adoption of AI in drug discovery.
Despite these challenges, the study emphasizes that AI’s integration with experimental methods - such as AI-guided high-throughput screening, in vitro and in vivo testing, and AI-assisted synthesis - has the potential to overcome traditional bottlenecks in drug development. By continuously refining AI methodologies and fostering collaboration between academia, industry, and regulatory bodies, AI-driven drug discovery could redefine the future of oncology treatments.
Future prospects and conclusion
The convergence of AI, big data, and computational biology is ushering in a new era of precision oncology. With AI’s ability to analyze multi-omics datasets and predict drug interactions with high accuracy, the pharmaceutical landscape is witnessing a shift towards more efficient, cost-effective, and patient-specific cancer treatments. As AI continues to evolve, its role in drug discovery will expand further, with advancements in quantum computing and multimodal AI offering even greater potential.
Overall, the study underscores AI’s transformative impact on oncology drug discovery. While challenges remain, the ongoing advancements in AI-driven methodologies hold the promise of significantly improving cancer treatment outcomes. By bridging the gap between computational power and experimental validation, AI is not only accelerating drug discovery but also making personalized medicine a tangible reality for cancer patients worldwide. | 5 |
The field of cancer treatment has long struggled with the immense costs and time-consuming nature of drug development. Traditional methods often take over a decade and billions of dollars to bring a single drug to market, with many compounds failing in late-stage trials due to efficacy or safety concerns. However, artificial intelligence (AI) is now revolutionizing this space by accelerating drug repurposing and designing new therapeutics with unprecedented speed and accuracy. The integration of AI in oncology drug discovery holds the promise of reducing development timelines, optimizing existing drugs, and unveiling novel treatment strategies.
A recent study titled Artificial Intelligence-Based Methods for Drug Repurposing and Development in Cancer, authored by Sara Herráiz-Gil, Elisa Nygren-Jiménez, Diana N. Acosta-Alonso, Carlos León, and Sara Guerrero-Aspizua, and published in Applied Sciences (2025, 15, 2798), presents an in-depth review of AI-driven techniques in drug discovery. The study highlights AI’s role in addressing critical challenges in oncology and explores the latest methodologies and applications in the field.
Role of AI in drug repurposing and new drug development
AI has emerged as a game-changer in the pharmaceutical industry, particularly in oncology, by offering powerful tools for drug repurposing and de novo drug design. Traditional drug repurposing involves identifying new uses for existing drugs, but AI significantly enhances this process by analyzing large-scale biological and chemical data. Machine learning and deep learning algorithms can predict drug-disease interactions, optimize drug efficacy, and minimize toxicity concerns.
The study discusses how knowledge graphs and neural networks are being employed to map complex relationships between drugs and diseases. Graph-based AI techniques allow researchers to identify potential drug candidates based on existing biological networks, while deep learning models can assess molecular interactions with remarkable precision. Generative AI models, such as reinforcement learning algorithms, are also gaining traction in de novo drug design, creating entirely new molecular structures optimized for cancer treatment. By leveraging multi-omics data, AI enables a more precise and personalized approach to therapy.
AI applications in cancer drug discovery
Several groundbreaking AI-driven projects have demonstrated the potential of this technology in oncology. The study outlines multiple case studies where AI was successfully applied to identify promising drug candidates. For instance, AI models have predicted potential therapies for chondrosarcoma, familial melanoma, and liver and lung cancers. By integrating diverse datasets, such as genomic profiles, protein interactions, and clinical trial results, these models provide insights into drug repositioning opportunities that might otherwise remain undiscovered.
Furthermore, AI has accelerated drug screening by predicting the binding affinity of molecules to cancer targets, reducing the need for costly and time-intensive laboratory testing. In hepatocellular carcinoma research, AI-driven de novo drug design has led to the development of a novel CDK20 inhibitor in record time, highlighting the efficiency of computational drug discovery. Additionally, AI-guided strategies have been instrumental in predicting patient responses to specific treatments, paving the way for more targeted and effective cancer therapies.
Experimental integration and challenges in AI-driven drug discovery
While AI presents remarkable opportunities, its integration into traditional experimental workflows comes with challenges. One major limitation is data quality - AI models rely on vast amounts of biomedical data, which can sometimes be inconsistent or biased. Ensuring data standardization and accuracy remains a critical challenge in training reliable AI models.
Another concern is the interpretability of AI predictions. Many deep learning models operate as “black boxes,” making it difficult for researchers to understand the rationale behind certain drug recommendations. To address this, explainable AI (XAI) techniques, such as SHAP and LIME, are being developed to enhance model transparency and regulatory acceptance. Ethical considerations, such as ensuring equitable access to AI-driven therapies and mitigating algorithmic biases, are also central to the responsible adoption of AI in drug discovery.
Despite these challenges, the study emphasizes that AI’s integration with experimental methods - such as AI-guided high-throughput screening, in vitro and in vivo testing, and AI-assisted synthesis - has the potential to overcome traditional bottlenecks in drug development. By continuously refining AI methodologies and fostering collaboration between academia, industry, and regulatory bodies, AI-driven drug discovery could redefine the future of oncology treatments.
Future prospects and conclusion
The convergence of AI, big data, and computational biology is ushering in a new era of precision oncology. With AI’s ability to analyze multi-omics datasets and predict drug interactions with high accuracy, the pharmaceutical landscape is witnessing a shift towards more efficient, cost-effective, and patient-specific cancer treatments. As AI continues to evolve, its role in drug discovery will expand further, with advancements in quantum computing and multimodal AI offering even greater potential.
Overall, the study underscores AI’s transformative impact on oncology drug discovery. While challenges remain, the ongoing advancements in AI-driven methodologies hold the promise of significantly improving cancer treatment outcomes. By bridging the gap between computational power and experimental validation, AI is not only accelerating drug discovery but also making personalized medicine a tangible reality for cancer patients worldwide. | 5 |
The field of cancer treatment has long struggled with the immense costs and time-consuming nature of drug development. Traditional methods often take over a decade and billions of dollars to bring a single drug to market, with many compounds failing in late-stage trials due to efficacy or safety concerns. However, artificial intelligence (AI) is now revolutionizing this space by accelerating drug repurposing and designing new therapeutics with unprecedented speed and accuracy. The integration of AI in oncology drug discovery holds the promise of reducing development timelines, optimizing existing drugs, and unveiling novel treatment strategies.
A recent study titled Artificial Intelligence-Based Methods for Drug Repurposing and Development in Cancer, authored by Sara Herráiz-Gil, Elisa Nygren-Jiménez, Diana N. Acosta-Alonso, Carlos León, and Sara Guerrero-Aspizua, and published in Applied Sciences (2025, 15, 2798), presents an in-depth review of AI-driven techniques in drug discovery. The study highlights AI’s role in addressing critical challenges in oncology and explores the latest methodologies and applications in the field.
Role of AI in drug repurposing and new drug development
AI has emerged as a game-changer in the pharmaceutical industry, particularly in oncology, by offering powerful tools for drug repurposing and de novo drug design. Traditional drug repurposing involves identifying new uses for existing drugs, but AI significantly enhances this process by analyzing large-scale biological and chemical data. Machine learning and deep learning algorithms can predict drug-disease interactions, optimize drug efficacy, and minimize toxicity concerns.
The study discusses how knowledge graphs and neural networks are being employed to map complex relationships between drugs and diseases. Graph-based AI techniques allow researchers to identify potential drug candidates based on existing biological networks, while deep learning models can assess molecular interactions with remarkable precision. Generative AI models, such as reinforcement learning algorithms, are also gaining traction in de novo drug design, creating entirely new molecular structures optimized for cancer treatment. By leveraging multi-omics data, AI enables a more precise and personalized approach to therapy.
AI applications in cancer drug discovery
Several groundbreaking AI-driven projects have demonstrated the potential of this technology in oncology. The study outlines multiple case studies where AI was successfully applied to identify promising drug candidates. For instance, AI models have predicted potential therapies for chondrosarcoma, familial melanoma, and liver and lung cancers. By integrating diverse datasets, such as genomic profiles, protein interactions, and clinical trial results, these models provide insights into drug repositioning opportunities that might otherwise remain undiscovered.
Furthermore, AI has accelerated drug screening by predicting the binding affinity of molecules to cancer targets, reducing the need for costly and time-intensive laboratory testing. In hepatocellular carcinoma research, AI-driven de novo drug design has led to the development of a novel CDK20 inhibitor in record time, highlighting the efficiency of computational drug discovery. Additionally, AI-guided strategies have been instrumental in predicting patient responses to specific treatments, paving the way for more targeted and effective cancer therapies.
Experimental integration and challenges in AI-driven drug discovery
While AI presents remarkable opportunities, its integration into traditional experimental workflows comes with challenges. One major limitation is data quality - AI models rely on vast amounts of biomedical data, which can sometimes be inconsistent or biased. Ensuring data standardization and accuracy remains a critical challenge in training reliable AI models.
Another concern is the interpretability of AI predictions. Many deep learning models operate as “black boxes,” making it difficult for researchers to understand the rationale behind certain drug recommendations. To address this, explainable AI (XAI) techniques, such as SHAP and LIME, are being developed to enhance model transparency and regulatory acceptance. Ethical considerations, such as ensuring equitable access to AI-driven therapies and mitigating algorithmic biases, are also central to the responsible adoption of AI in drug discovery.
Despite these challenges, the study emphasizes that AI’s integration with experimental methods - such as AI-guided high-throughput screening, in vitro and in vivo testing, and AI-assisted synthesis - has the potential to overcome traditional bottlenecks in drug development. By continuously refining AI methodologies and fostering collaboration between academia, industry, and regulatory bodies, AI-driven drug discovery could redefine the future of oncology treatments.
Future prospects and conclusion
The convergence of AI, big data, and computational biology is ushering in a new era of precision oncology. With AI’s ability to analyze multi-omics datasets and predict drug interactions with high accuracy, the pharmaceutical landscape is witnessing a shift towards more efficient, cost-effective, and patient-specific cancer treatments. As AI continues to evolve, its role in drug discovery will expand further, with advancements in quantum computing and multimodal AI offering even greater potential.
Overall, the study underscores AI’s transformative impact on oncology drug discovery. While challenges remain, the ongoing advancements in AI-driven methodologies hold the promise of significantly improving cancer treatment outcomes. By bridging the gap between computational power and experimental validation, AI is not only accelerating drug discovery but also making personalized medicine a tangible reality for cancer patients worldwide. | 5 |
The field of cancer treatment has long struggled with the immense costs and time-consuming nature of drug development. Traditional methods often take over a decade and billions of dollars to bring a single drug to market, with many compounds failing in late-stage trials due to efficacy or safety concerns. However, artificial intelligence (AI) is now revolutionizing this space by accelerating drug repurposing and designing new therapeutics with unprecedented speed and accuracy. The integration of AI in oncology drug discovery holds the promise of reducing development timelines, optimizing existing drugs, and unveiling novel treatment strategies.
A recent study titled Artificial Intelligence-Based Methods for Drug Repurposing and Development in Cancer, authored by Sara Herráiz-Gil, Elisa Nygren-Jiménez, Diana N. Acosta-Alonso, Carlos León, and Sara Guerrero-Aspizua, and published in Applied Sciences (2025, 15, 2798), presents an in-depth review of AI-driven techniques in drug discovery. The study highlights AI’s role in addressing critical challenges in oncology and explores the latest methodologies and applications in the field.
Role of AI in drug repurposing and new drug development
AI has emerged as a game-changer in the pharmaceutical industry, particularly in oncology, by offering powerful tools for drug repurposing and de novo drug design. Traditional drug repurposing involves identifying new uses for existing drugs, but AI significantly enhances this process by analyzing large-scale biological and chemical data. Machine learning and deep learning algorithms can predict drug-disease interactions, optimize drug efficacy, and minimize toxicity concerns.
The study discusses how knowledge graphs and neural networks are being employed to map complex relationships between drugs and diseases. Graph-based AI techniques allow researchers to identify potential drug candidates based on existing biological networks, while deep learning models can assess molecular interactions with remarkable precision. Generative AI models, such as reinforcement learning algorithms, are also gaining traction in de novo drug design, creating entirely new molecular structures optimized for cancer treatment. By leveraging multi-omics data, AI enables a more precise and personalized approach to therapy.
AI applications in cancer drug discovery
Several groundbreaking AI-driven projects have demonstrated the potential of this technology in oncology. The study outlines multiple case studies where AI was successfully applied to identify promising drug candidates. For instance, AI models have predicted potential therapies for chondrosarcoma, familial melanoma, and liver and lung cancers. By integrating diverse datasets, such as genomic profiles, protein interactions, and clinical trial results, these models provide insights into drug repositioning opportunities that might otherwise remain undiscovered.
Furthermore, AI has accelerated drug screening by predicting the binding affinity of molecules to cancer targets, reducing the need for costly and time-intensive laboratory testing. In hepatocellular carcinoma research, AI-driven de novo drug design has led to the development of a novel CDK20 inhibitor in record time, highlighting the efficiency of computational drug discovery. Additionally, AI-guided strategies have been instrumental in predicting patient responses to specific treatments, paving the way for more targeted and effective cancer therapies.
Experimental integration and challenges in AI-driven drug discovery
While AI presents remarkable opportunities, its integration into traditional experimental workflows comes with challenges. One major limitation is data quality - AI models rely on vast amounts of biomedical data, which can sometimes be inconsistent or biased. Ensuring data standardization and accuracy remains a critical challenge in training reliable AI models.
Another concern is the interpretability of AI predictions. Many deep learning models operate as “black boxes,” making it difficult for researchers to understand the rationale behind certain drug recommendations. To address this, explainable AI (XAI) techniques, such as SHAP and LIME, are being developed to enhance model transparency and regulatory acceptance. Ethical considerations, such as ensuring equitable access to AI-driven therapies and mitigating algorithmic biases, are also central to the responsible adoption of AI in drug discovery.
Despite these challenges, the study emphasizes that AI’s integration with experimental methods - such as AI-guided high-throughput screening, in vitro and in vivo testing, and AI-assisted synthesis - has the potential to overcome traditional bottlenecks in drug development. By continuously refining AI methodologies and fostering collaboration between academia, industry, and regulatory bodies, AI-driven drug discovery could redefine the future of oncology treatments.
Future prospects and conclusion
The convergence of AI, big data, and computational biology is ushering in a new era of precision oncology. With AI’s ability to analyze multi-omics datasets and predict drug interactions with high accuracy, the pharmaceutical landscape is witnessing a shift towards more efficient, cost-effective, and patient-specific cancer treatments. As AI continues to evolve, its role in drug discovery will expand further, with advancements in quantum computing and multimodal AI offering even greater potential.
Overall, the study underscores AI’s transformative impact on oncology drug discovery. While challenges remain, the ongoing advancements in AI-driven methodologies hold the promise of significantly improving cancer treatment outcomes. By bridging the gap between computational power and experimental validation, AI is not only accelerating drug discovery but also making personalized medicine a tangible reality for cancer patients worldwide. | 5 |
The field of cancer treatment has long struggled with the immense costs and time-consuming nature of drug development. Traditional methods often take over a decade and billions of dollars to bring a single drug to market, with many compounds failing in late-stage trials due to efficacy or safety concerns. However, artificial intelligence (AI) is now revolutionizing this space by accelerating drug repurposing and designing new therapeutics with unprecedented speed and accuracy. The integration of AI in oncology drug discovery holds the promise of reducing development timelines, optimizing existing drugs, and unveiling novel treatment strategies.
A recent study titled Artificial Intelligence-Based Methods for Drug Repurposing and Development in Cancer, authored by Sara Herráiz-Gil, Elisa Nygren-Jiménez, Diana N. Acosta-Alonso, Carlos León, and Sara Guerrero-Aspizua, and published in Applied Sciences (2025, 15, 2798), presents an in-depth review of AI-driven techniques in drug discovery. The study highlights AI’s role in addressing critical challenges in oncology and explores the latest methodologies and applications in the field.
Role of AI in drug repurposing and new drug development
AI has emerged as a game-changer in the pharmaceutical industry, particularly in oncology, by offering powerful tools for drug repurposing and de novo drug design. Traditional drug repurposing involves identifying new uses for existing drugs, but AI significantly enhances this process by analyzing large-scale biological and chemical data. Machine learning and deep learning algorithms can predict drug-disease interactions, optimize drug efficacy, and minimize toxicity concerns.
The study discusses how knowledge graphs and neural networks are being employed to map complex relationships between drugs and diseases. Graph-based AI techniques allow researchers to identify potential drug candidates based on existing biological networks, while deep learning models can assess molecular interactions with remarkable precision. Generative AI models, such as reinforcement learning algorithms, are also gaining traction in de novo drug design, creating entirely new molecular structures optimized for cancer treatment. By leveraging multi-omics data, AI enables a more precise and personalized approach to therapy.
AI applications in cancer drug discovery
Several groundbreaking AI-driven projects have demonstrated the potential of this technology in oncology. The study outlines multiple case studies where AI was successfully applied to identify promising drug candidates. For instance, AI models have predicted potential therapies for chondrosarcoma, familial melanoma, and liver and lung cancers. By integrating diverse datasets, such as genomic profiles, protein interactions, and clinical trial results, these models provide insights into drug repositioning opportunities that might otherwise remain undiscovered.
Furthermore, AI has accelerated drug screening by predicting the binding affinity of molecules to cancer targets, reducing the need for costly and time-intensive laboratory testing. In hepatocellular carcinoma research, AI-driven de novo drug design has led to the development of a novel CDK20 inhibitor in record time, highlighting the efficiency of computational drug discovery. Additionally, AI-guided strategies have been instrumental in predicting patient responses to specific treatments, paving the way for more targeted and effective cancer therapies.
Experimental integration and challenges in AI-driven drug discovery
While AI presents remarkable opportunities, its integration into traditional experimental workflows comes with challenges. One major limitation is data quality - AI models rely on vast amounts of biomedical data, which can sometimes be inconsistent or biased. Ensuring data standardization and accuracy remains a critical challenge in training reliable AI models.
Another concern is the interpretability of AI predictions. Many deep learning models operate as “black boxes,” making it difficult for researchers to understand the rationale behind certain drug recommendations. To address this, explainable AI (XAI) techniques, such as SHAP and LIME, are being developed to enhance model transparency and regulatory acceptance. Ethical considerations, such as ensuring equitable access to AI-driven therapies and mitigating algorithmic biases, are also central to the responsible adoption of AI in drug discovery.
Despite these challenges, the study emphasizes that AI’s integration with experimental methods - such as AI-guided high-throughput screening, in vitro and in vivo testing, and AI-assisted synthesis - has the potential to overcome traditional bottlenecks in drug development. By continuously refining AI methodologies and fostering collaboration between academia, industry, and regulatory bodies, AI-driven drug discovery could redefine the future of oncology treatments.
Future prospects and conclusion
The convergence of AI, big data, and computational biology is ushering in a new era of precision oncology. With AI’s ability to analyze multi-omics datasets and predict drug interactions with high accuracy, the pharmaceutical landscape is witnessing a shift towards more efficient, cost-effective, and patient-specific cancer treatments. As AI continues to evolve, its role in drug discovery will expand further, with advancements in quantum computing and multimodal AI offering even greater potential.
Overall, the study underscores AI’s transformative impact on oncology drug discovery. While challenges remain, the ongoing advancements in AI-driven methodologies hold the promise of significantly improving cancer treatment outcomes. By bridging the gap between computational power and experimental validation, AI is not only accelerating drug discovery but also making personalized medicine a tangible reality for cancer patients worldwide. | 5 |
The United States has implemented new AI export rules on Monday. These rules control the export of artificial intelligence (AI) technology. The aim is to safeguard national security. This also seeks to maintain a competitive edge. These regulations impact various sectors. They affect tech companies and research institutions alike. The changes are significant.
Key Aspects of the Regulations
The rules create three tiers of countries for exports of advanced AI chips and technology. Tier 1 includes close allies and partners such as Australia, Japan, South Korea, and Taiwan, who will face no restrictions. Tier 2 countries will face caps on the number of AI chips they can import, while Tier 3 countries will require a license for any exports. China, Russia etc are in tier 3 countries.
The rules affect the export of AI software. They also target hardware. Certain types of chips are included. High-performance computing resources are also impacted. The regulations require licenses for exports. These are needed to certain countries. This process can be complex. Companies must prove their exports are not a risk. They must demonstrate responsible use.
Which Countries Are Affected
However the government has not openly told the public about limitations for a particular country, but surely targets certain nations. China is a major focus. Russia is also under scrutiny. Other countries with security concerns are also listed. The regulations are not uniform.
They vary based on the destination country. Some nations face strict restrictions. Others may have less stringent requirements. The US is adapting these rules constantly. This is due to the rapid evolution of AI. 20 countries are listed in tier 1 where the government will not ask any license and further documentation.
The second priority will be given to USA supported countries including Saudi Arabia, Israel and other 120 countries. The most restrictions will be faced by China and Russia where they need to complete proper documentation and licenses to export or import AI chips from USA.
Impact on Businesses
These export rules create new challenges for companies. Tech firms must now navigate complex regulations. They have to obtain export licenses. Compliance can be expensive. It can also be time-consuming. There might be delays in shipments. This can disrupt supply chains.
It can also affect partnerships. Businesses might need to re-evaluate their global strategies. They may need to consider alternative markets. The rules also affect research institutions. Collaborations with international partners require careful review.
Impact on Research and Development
These rules can slow down AI research. International collaboration is crucial for AI advancement. These collaborations are facing new hurdles. Researchers need to comply with export regulations.
This can limit the exchange of information. It can hinder the sharing of cutting-edge research. Some worry this could stall global AI progress. The US government says safeguards are necessary. They aim to balance innovation with national security.
Debate and Criticisms
These new rules have sparked debate. Some argue they are necessary for national security. They feel the risk of misuse is real. Others express concern over limitations. They worry that regulations could stifle innovation. They also worry about economic competitiveness.
There is also worry over unintended consequences. The regulations could create barriers for legitimate businesses. Some are concerned about the scope of the restrictions. The line between national security concerns and trade protectionism is debated.
Why the New Rules
Several factors triggered these rules. Rapid AI advancements are one reason. AI has become very powerful. It has applications in many fields. Some applications are military in nature. The US wants to control the technology. It seeks to prevent misuse. There are concerns about adversaries. These concerns relate to potential threats. The US also wants to protect its innovation. It aims to avoid technology transfers that could weaken its position.
Future Outlook
The AI export rules are expected to evolve. The US government will continue to adjust. These changes will happen based on tech developments. They will also react to global changes. Ongoing dialogue is needed. It must involve stakeholders in the industry.
This will help navigate the new landscape. There needs to be balance. It must be between security and innovation. The aim is to ensure responsible AI development and deployment. The future of AI exports remains uncertain. It will depend on further changes in the tech field. | 5 |
The United States has implemented new AI export rules on Monday. These rules control the export of artificial intelligence (AI) technology. The aim is to safeguard national security. This also seeks to maintain a competitive edge. These regulations impact various sectors. They affect tech companies and research institutions alike. The changes are significant.
Key Aspects of the Regulations
The rules create three tiers of countries for exports of advanced AI chips and technology. Tier 1 includes close allies and partners such as Australia, Japan, South Korea, and Taiwan, who will face no restrictions. Tier 2 countries will face caps on the number of AI chips they can import, while Tier 3 countries will require a license for any exports. China, Russia etc are in tier 3 countries.
The rules affect the export of AI software. They also target hardware. Certain types of chips are included. High-performance computing resources are also impacted. The regulations require licenses for exports. These are needed to certain countries. This process can be complex. Companies must prove their exports are not a risk. They must demonstrate responsible use.
Which Countries Are Affected
However the government has not openly told the public about limitations for a particular country, but surely targets certain nations. China is a major focus. Russia is also under scrutiny. Other countries with security concerns are also listed. The regulations are not uniform.
They vary based on the destination country. Some nations face strict restrictions. Others may have less stringent requirements. The US is adapting these rules constantly. This is due to the rapid evolution of AI. 20 countries are listed in tier 1 where the government will not ask any license and further documentation.
The second priority will be given to USA supported countries including Saudi Arabia, Israel and other 120 countries. The most restrictions will be faced by China and Russia where they need to complete proper documentation and licenses to export or import AI chips from USA.
Impact on Businesses
These export rules create new challenges for companies. Tech firms must now navigate complex regulations. They have to obtain export licenses. Compliance can be expensive. It can also be time-consuming. There might be delays in shipments. This can disrupt supply chains.
It can also affect partnerships. Businesses might need to re-evaluate their global strategies. They may need to consider alternative markets. The rules also affect research institutions. Collaborations with international partners require careful review.
Impact on Research and Development
These rules can slow down AI research. International collaboration is crucial for AI advancement. These collaborations are facing new hurdles. Researchers need to comply with export regulations.
This can limit the exchange of information. It can hinder the sharing of cutting-edge research. Some worry this could stall global AI progress. The US government says safeguards are necessary. They aim to balance innovation with national security.
Debate and Criticisms
These new rules have sparked debate. Some argue they are necessary for national security. They feel the risk of misuse is real. Others express concern over limitations. They worry that regulations could stifle innovation. They also worry about economic competitiveness.
There is also worry over unintended consequences. The regulations could create barriers for legitimate businesses. Some are concerned about the scope of the restrictions. The line between national security concerns and trade protectionism is debated.
Why the New Rules
Several factors triggered these rules. Rapid AI advancements are one reason. AI has become very powerful. It has applications in many fields. Some applications are military in nature. The US wants to control the technology. It seeks to prevent misuse. There are concerns about adversaries. These concerns relate to potential threats. The US also wants to protect its innovation. It aims to avoid technology transfers that could weaken its position.
Future Outlook
The AI export rules are expected to evolve. The US government will continue to adjust. These changes will happen based on tech developments. They will also react to global changes. Ongoing dialogue is needed. It must involve stakeholders in the industry.
This will help navigate the new landscape. There needs to be balance. It must be between security and innovation. The aim is to ensure responsible AI development and deployment. The future of AI exports remains uncertain. It will depend on further changes in the tech field. | 5 |
The United States has implemented new AI export rules on Monday. These rules control the export of artificial intelligence (AI) technology. The aim is to safeguard national security. This also seeks to maintain a competitive edge. These regulations impact various sectors. They affect tech companies and research institutions alike. The changes are significant.
Key Aspects of the Regulations
The rules create three tiers of countries for exports of advanced AI chips and technology. Tier 1 includes close allies and partners such as Australia, Japan, South Korea, and Taiwan, who will face no restrictions. Tier 2 countries will face caps on the number of AI chips they can import, while Tier 3 countries will require a license for any exports. China, Russia etc are in tier 3 countries.
The rules affect the export of AI software. They also target hardware. Certain types of chips are included. High-performance computing resources are also impacted. The regulations require licenses for exports. These are needed to certain countries. This process can be complex. Companies must prove their exports are not a risk. They must demonstrate responsible use.
Which Countries Are Affected
However the government has not openly told the public about limitations for a particular country, but surely targets certain nations. China is a major focus. Russia is also under scrutiny. Other countries with security concerns are also listed. The regulations are not uniform.
They vary based on the destination country. Some nations face strict restrictions. Others may have less stringent requirements. The US is adapting these rules constantly. This is due to the rapid evolution of AI. 20 countries are listed in tier 1 where the government will not ask any license and further documentation.
The second priority will be given to USA supported countries including Saudi Arabia, Israel and other 120 countries. The most restrictions will be faced by China and Russia where they need to complete proper documentation and licenses to export or import AI chips from USA.
Impact on Businesses
These export rules create new challenges for companies. Tech firms must now navigate complex regulations. They have to obtain export licenses. Compliance can be expensive. It can also be time-consuming. There might be delays in shipments. This can disrupt supply chains.
It can also affect partnerships. Businesses might need to re-evaluate their global strategies. They may need to consider alternative markets. The rules also affect research institutions. Collaborations with international partners require careful review.
Impact on Research and Development
These rules can slow down AI research. International collaboration is crucial for AI advancement. These collaborations are facing new hurdles. Researchers need to comply with export regulations.
This can limit the exchange of information. It can hinder the sharing of cutting-edge research. Some worry this could stall global AI progress. The US government says safeguards are necessary. They aim to balance innovation with national security.
Debate and Criticisms
These new rules have sparked debate. Some argue they are necessary for national security. They feel the risk of misuse is real. Others express concern over limitations. They worry that regulations could stifle innovation. They also worry about economic competitiveness.
There is also worry over unintended consequences. The regulations could create barriers for legitimate businesses. Some are concerned about the scope of the restrictions. The line between national security concerns and trade protectionism is debated.
Why the New Rules
Several factors triggered these rules. Rapid AI advancements are one reason. AI has become very powerful. It has applications in many fields. Some applications are military in nature. The US wants to control the technology. It seeks to prevent misuse. There are concerns about adversaries. These concerns relate to potential threats. The US also wants to protect its innovation. It aims to avoid technology transfers that could weaken its position.
Future Outlook
The AI export rules are expected to evolve. The US government will continue to adjust. These changes will happen based on tech developments. They will also react to global changes. Ongoing dialogue is needed. It must involve stakeholders in the industry.
This will help navigate the new landscape. There needs to be balance. It must be between security and innovation. The aim is to ensure responsible AI development and deployment. The future of AI exports remains uncertain. It will depend on further changes in the tech field. | 5 |
The United States has implemented new AI export rules on Monday. These rules control the export of artificial intelligence (AI) technology. The aim is to safeguard national security. This also seeks to maintain a competitive edge. These regulations impact various sectors. They affect tech companies and research institutions alike. The changes are significant.
Key Aspects of the Regulations
The rules create three tiers of countries for exports of advanced AI chips and technology. Tier 1 includes close allies and partners such as Australia, Japan, South Korea, and Taiwan, who will face no restrictions. Tier 2 countries will face caps on the number of AI chips they can import, while Tier 3 countries will require a license for any exports. China, Russia etc are in tier 3 countries.
The rules affect the export of AI software. They also target hardware. Certain types of chips are included. High-performance computing resources are also impacted. The regulations require licenses for exports. These are needed to certain countries. This process can be complex. Companies must prove their exports are not a risk. They must demonstrate responsible use.
Which Countries Are Affected
However the government has not openly told the public about limitations for a particular country, but surely targets certain nations. China is a major focus. Russia is also under scrutiny. Other countries with security concerns are also listed. The regulations are not uniform.
They vary based on the destination country. Some nations face strict restrictions. Others may have less stringent requirements. The US is adapting these rules constantly. This is due to the rapid evolution of AI. 20 countries are listed in tier 1 where the government will not ask any license and further documentation.
The second priority will be given to USA supported countries including Saudi Arabia, Israel and other 120 countries. The most restrictions will be faced by China and Russia where they need to complete proper documentation and licenses to export or import AI chips from USA.
Impact on Businesses
These export rules create new challenges for companies. Tech firms must now navigate complex regulations. They have to obtain export licenses. Compliance can be expensive. It can also be time-consuming. There might be delays in shipments. This can disrupt supply chains.
It can also affect partnerships. Businesses might need to re-evaluate their global strategies. They may need to consider alternative markets. The rules also affect research institutions. Collaborations with international partners require careful review.
Impact on Research and Development
These rules can slow down AI research. International collaboration is crucial for AI advancement. These collaborations are facing new hurdles. Researchers need to comply with export regulations.
This can limit the exchange of information. It can hinder the sharing of cutting-edge research. Some worry this could stall global AI progress. The US government says safeguards are necessary. They aim to balance innovation with national security.
Debate and Criticisms
These new rules have sparked debate. Some argue they are necessary for national security. They feel the risk of misuse is real. Others express concern over limitations. They worry that regulations could stifle innovation. They also worry about economic competitiveness.
There is also worry over unintended consequences. The regulations could create barriers for legitimate businesses. Some are concerned about the scope of the restrictions. The line between national security concerns and trade protectionism is debated.
Why the New Rules
Several factors triggered these rules. Rapid AI advancements are one reason. AI has become very powerful. It has applications in many fields. Some applications are military in nature. The US wants to control the technology. It seeks to prevent misuse. There are concerns about adversaries. These concerns relate to potential threats. The US also wants to protect its innovation. It aims to avoid technology transfers that could weaken its position.
Future Outlook
The AI export rules are expected to evolve. The US government will continue to adjust. These changes will happen based on tech developments. They will also react to global changes. Ongoing dialogue is needed. It must involve stakeholders in the industry.
This will help navigate the new landscape. There needs to be balance. It must be between security and innovation. The aim is to ensure responsible AI development and deployment. The future of AI exports remains uncertain. It will depend on further changes in the tech field. | 5 |
The United States has implemented new AI export rules on Monday. These rules control the export of artificial intelligence (AI) technology. The aim is to safeguard national security. This also seeks to maintain a competitive edge. These regulations impact various sectors. They affect tech companies and research institutions alike. The changes are significant.
Key Aspects of the Regulations
The rules create three tiers of countries for exports of advanced AI chips and technology. Tier 1 includes close allies and partners such as Australia, Japan, South Korea, and Taiwan, who will face no restrictions. Tier 2 countries will face caps on the number of AI chips they can import, while Tier 3 countries will require a license for any exports. China, Russia etc are in tier 3 countries.
The rules affect the export of AI software. They also target hardware. Certain types of chips are included. High-performance computing resources are also impacted. The regulations require licenses for exports. These are needed to certain countries. This process can be complex. Companies must prove their exports are not a risk. They must demonstrate responsible use.
Which Countries Are Affected
However the government has not openly told the public about limitations for a particular country, but surely targets certain nations. China is a major focus. Russia is also under scrutiny. Other countries with security concerns are also listed. The regulations are not uniform.
They vary based on the destination country. Some nations face strict restrictions. Others may have less stringent requirements. The US is adapting these rules constantly. This is due to the rapid evolution of AI. 20 countries are listed in tier 1 where the government will not ask any license and further documentation.
The second priority will be given to USA supported countries including Saudi Arabia, Israel and other 120 countries. The most restrictions will be faced by China and Russia where they need to complete proper documentation and licenses to export or import AI chips from USA.
Impact on Businesses
These export rules create new challenges for companies. Tech firms must now navigate complex regulations. They have to obtain export licenses. Compliance can be expensive. It can also be time-consuming. There might be delays in shipments. This can disrupt supply chains.
It can also affect partnerships. Businesses might need to re-evaluate their global strategies. They may need to consider alternative markets. The rules also affect research institutions. Collaborations with international partners require careful review.
Impact on Research and Development
These rules can slow down AI research. International collaboration is crucial for AI advancement. These collaborations are facing new hurdles. Researchers need to comply with export regulations.
This can limit the exchange of information. It can hinder the sharing of cutting-edge research. Some worry this could stall global AI progress. The US government says safeguards are necessary. They aim to balance innovation with national security.
Debate and Criticisms
These new rules have sparked debate. Some argue they are necessary for national security. They feel the risk of misuse is real. Others express concern over limitations. They worry that regulations could stifle innovation. They also worry about economic competitiveness.
There is also worry over unintended consequences. The regulations could create barriers for legitimate businesses. Some are concerned about the scope of the restrictions. The line between national security concerns and trade protectionism is debated.
Why the New Rules
Several factors triggered these rules. Rapid AI advancements are one reason. AI has become very powerful. It has applications in many fields. Some applications are military in nature. The US wants to control the technology. It seeks to prevent misuse. There are concerns about adversaries. These concerns relate to potential threats. The US also wants to protect its innovation. It aims to avoid technology transfers that could weaken its position.
Future Outlook
The AI export rules are expected to evolve. The US government will continue to adjust. These changes will happen based on tech developments. They will also react to global changes. Ongoing dialogue is needed. It must involve stakeholders in the industry.
This will help navigate the new landscape. There needs to be balance. It must be between security and innovation. The aim is to ensure responsible AI development and deployment. The future of AI exports remains uncertain. It will depend on further changes in the tech field. | 5 |
In an age where complex societal challenges exceed human capabilities, the integration of artificial intelligence (AI) into collective decision-making has emerged as a transformative force. While some fear AI replacing human intellect, a groundbreaking study argues that AI, when integrated effectively, can enhance human collective intelligence rather than replace it.
The study, titled AI-enhanced Collective Intelligence, authored by Hao Cui and Taha Yasseri and published in Patterns, explores the synergy between human cognition and AI, advocating for a hybrid model where AI augments human capabilities to tackle increasingly intricate problems.
The evolution of collective intelligence
Collective intelligence (CI) has long been recognized as the superior outcome of collaborative human efforts. From ancient social structures to modern technological advancements, humanity has leveraged collective knowledge to drive progress. The Internet and large-scale collaboration platforms, such as Wikipedia and crowdsourced research projects, have already elevated CI by enabling knowledge sharing across vast networks.
However, the advent of AI introduces new dimensions to CI. Unlike traditional models, where human collectives function independently, AI-enhanced CI proposes an interconnected system where AI and human agents coalesce to form superior decision-making units. AI brings computational efficiency, pattern recognition, and data-processing prowess, while humans contribute intuition, creativity, and ethical considerations. Together, these complementary capabilities can yield unprecedented levels of intelligence and problem-solving efficacy.
Understanding the human-AI hybrid model
The study presents a multilayer representation of AI-enhanced CI, structured around three core layers: cognition, physical, and information. In this model, humans and AI interact dynamically within and across these layers, forming an interconnected system that amplifies collective problem-solving potential.
Cognition Layer: This includes the mental processes of both humans and AI, encompassing decision-making, sense-making, and reasoning. AI can enhance cognition by providing vast amounts of structured information, aiding in knowledge synthesis. Physical Layer: This represents tangible interactions between humans and AI, such as AI-driven robotic automation in industries or real-world collaboration in diagnostics and engineering. Information Layer: This layer governs the exchange of data between human and AI agents, determining how insights are shared and integrated to enhance CI.
By conceptualizing CI as a networked system, the study highlights how AI’s consistency, scalability, and adaptability can complement human intelligence, reducing biases and inconsistencies often present in human decision-making.
Applications and real-world impact
The real-world implications of AI-enhanced CI span multiple domains, from scientific research to policy-making, healthcare, and environmental sustainability. The study examines cases where AI-driven CI has already made significant impacts:
Medical Diagnostics : AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making.
: AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making. Misinformation Detection : In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time.
: In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time. Environmental Conservation: Projects such as Litterati and eBird leverage AI-enhanced CI to engage communities in identifying pollution patterns and monitoring biodiversity, respectively. AI analyzes large datasets generated by human contributors, leading to more effective conservation strategies.
These examples illustrate the potential of AI-enhanced CI to drive efficiency, accuracy, and scalability in diverse fields, offering innovative solutions to global challenges.
Challenges and future directions
Despite its promise, AI-enhanced CI faces several challenges. The study underscores concerns related to AI bias, transparency, and ethical considerations. AI models are only as good as the data they are trained on, and biases in data can lead to flawed decision-making. Additionally, the lack of explainability in AI-generated outcomes raises questions about accountability and trust in human-AI collaborations.
Another challenge is maintaining human motivation and engagement in AI-integrated decision-making systems. Over-reliance on AI could lead to diminished human participation, reducing the benefits of diverse perspectives and ethical reasoning that human collectives bring to the table.
The future of AI-enhanced CI lies in refining human-AI interaction frameworks to ensure optimal collaboration. Researchers call for the development of AI systems that are transparent, interpretable, and aligned with human values. Moreover, interdisciplinary cooperation between cognitive scientists, ethicists, and AI developers is necessary to design CI systems that balance computational efficiency with ethical and social considerations.
Conclusion: The path forward
The study by Cui and Yasseri presents a compelling vision for the future of collective intelligence, where AI is not a replacement for human intellect but a catalyst for enhanced decision-making. By fostering a symbiotic relationship between AI and human cognition, society can unlock new levels of problem-solving capabilities, addressing global challenges with greater efficiency and precision.
As AI continues to evolve, the focus must shift toward creating hybrid intelligence systems that integrate AI’s strengths while preserving human values, creativity, and ethical judgment. The ultimate goal is not to substitute human intelligence but to augment it, ensuring that the combined power of AI and human collectives drives progress in a rapidly changing world. | 5 |
In an age where complex societal challenges exceed human capabilities, the integration of artificial intelligence (AI) into collective decision-making has emerged as a transformative force. While some fear AI replacing human intellect, a groundbreaking study argues that AI, when integrated effectively, can enhance human collective intelligence rather than replace it.
The study, titled AI-enhanced Collective Intelligence, authored by Hao Cui and Taha Yasseri and published in Patterns, explores the synergy between human cognition and AI, advocating for a hybrid model where AI augments human capabilities to tackle increasingly intricate problems.
The evolution of collective intelligence
Collective intelligence (CI) has long been recognized as the superior outcome of collaborative human efforts. From ancient social structures to modern technological advancements, humanity has leveraged collective knowledge to drive progress. The Internet and large-scale collaboration platforms, such as Wikipedia and crowdsourced research projects, have already elevated CI by enabling knowledge sharing across vast networks.
However, the advent of AI introduces new dimensions to CI. Unlike traditional models, where human collectives function independently, AI-enhanced CI proposes an interconnected system where AI and human agents coalesce to form superior decision-making units. AI brings computational efficiency, pattern recognition, and data-processing prowess, while humans contribute intuition, creativity, and ethical considerations. Together, these complementary capabilities can yield unprecedented levels of intelligence and problem-solving efficacy.
Understanding the human-AI hybrid model
The study presents a multilayer representation of AI-enhanced CI, structured around three core layers: cognition, physical, and information. In this model, humans and AI interact dynamically within and across these layers, forming an interconnected system that amplifies collective problem-solving potential.
Cognition Layer: This includes the mental processes of both humans and AI, encompassing decision-making, sense-making, and reasoning. AI can enhance cognition by providing vast amounts of structured information, aiding in knowledge synthesis. Physical Layer: This represents tangible interactions between humans and AI, such as AI-driven robotic automation in industries or real-world collaboration in diagnostics and engineering. Information Layer: This layer governs the exchange of data between human and AI agents, determining how insights are shared and integrated to enhance CI.
By conceptualizing CI as a networked system, the study highlights how AI’s consistency, scalability, and adaptability can complement human intelligence, reducing biases and inconsistencies often present in human decision-making.
Applications and real-world impact
The real-world implications of AI-enhanced CI span multiple domains, from scientific research to policy-making, healthcare, and environmental sustainability. The study examines cases where AI-driven CI has already made significant impacts:
Medical Diagnostics : AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making.
: AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making. Misinformation Detection : In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time.
: In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time. Environmental Conservation: Projects such as Litterati and eBird leverage AI-enhanced CI to engage communities in identifying pollution patterns and monitoring biodiversity, respectively. AI analyzes large datasets generated by human contributors, leading to more effective conservation strategies.
These examples illustrate the potential of AI-enhanced CI to drive efficiency, accuracy, and scalability in diverse fields, offering innovative solutions to global challenges.
Challenges and future directions
Despite its promise, AI-enhanced CI faces several challenges. The study underscores concerns related to AI bias, transparency, and ethical considerations. AI models are only as good as the data they are trained on, and biases in data can lead to flawed decision-making. Additionally, the lack of explainability in AI-generated outcomes raises questions about accountability and trust in human-AI collaborations.
Another challenge is maintaining human motivation and engagement in AI-integrated decision-making systems. Over-reliance on AI could lead to diminished human participation, reducing the benefits of diverse perspectives and ethical reasoning that human collectives bring to the table.
The future of AI-enhanced CI lies in refining human-AI interaction frameworks to ensure optimal collaboration. Researchers call for the development of AI systems that are transparent, interpretable, and aligned with human values. Moreover, interdisciplinary cooperation between cognitive scientists, ethicists, and AI developers is necessary to design CI systems that balance computational efficiency with ethical and social considerations.
Conclusion: The path forward
The study by Cui and Yasseri presents a compelling vision for the future of collective intelligence, where AI is not a replacement for human intellect but a catalyst for enhanced decision-making. By fostering a symbiotic relationship between AI and human cognition, society can unlock new levels of problem-solving capabilities, addressing global challenges with greater efficiency and precision.
As AI continues to evolve, the focus must shift toward creating hybrid intelligence systems that integrate AI’s strengths while preserving human values, creativity, and ethical judgment. The ultimate goal is not to substitute human intelligence but to augment it, ensuring that the combined power of AI and human collectives drives progress in a rapidly changing world. | 5 |
In an age where complex societal challenges exceed human capabilities, the integration of artificial intelligence (AI) into collective decision-making has emerged as a transformative force. While some fear AI replacing human intellect, a groundbreaking study argues that AI, when integrated effectively, can enhance human collective intelligence rather than replace it.
The study, titled AI-enhanced Collective Intelligence, authored by Hao Cui and Taha Yasseri and published in Patterns, explores the synergy between human cognition and AI, advocating for a hybrid model where AI augments human capabilities to tackle increasingly intricate problems.
The evolution of collective intelligence
Collective intelligence (CI) has long been recognized as the superior outcome of collaborative human efforts. From ancient social structures to modern technological advancements, humanity has leveraged collective knowledge to drive progress. The Internet and large-scale collaboration platforms, such as Wikipedia and crowdsourced research projects, have already elevated CI by enabling knowledge sharing across vast networks.
However, the advent of AI introduces new dimensions to CI. Unlike traditional models, where human collectives function independently, AI-enhanced CI proposes an interconnected system where AI and human agents coalesce to form superior decision-making units. AI brings computational efficiency, pattern recognition, and data-processing prowess, while humans contribute intuition, creativity, and ethical considerations. Together, these complementary capabilities can yield unprecedented levels of intelligence and problem-solving efficacy.
Understanding the human-AI hybrid model
The study presents a multilayer representation of AI-enhanced CI, structured around three core layers: cognition, physical, and information. In this model, humans and AI interact dynamically within and across these layers, forming an interconnected system that amplifies collective problem-solving potential.
Cognition Layer: This includes the mental processes of both humans and AI, encompassing decision-making, sense-making, and reasoning. AI can enhance cognition by providing vast amounts of structured information, aiding in knowledge synthesis. Physical Layer: This represents tangible interactions between humans and AI, such as AI-driven robotic automation in industries or real-world collaboration in diagnostics and engineering. Information Layer: This layer governs the exchange of data between human and AI agents, determining how insights are shared and integrated to enhance CI.
By conceptualizing CI as a networked system, the study highlights how AI’s consistency, scalability, and adaptability can complement human intelligence, reducing biases and inconsistencies often present in human decision-making.
Applications and real-world impact
The real-world implications of AI-enhanced CI span multiple domains, from scientific research to policy-making, healthcare, and environmental sustainability. The study examines cases where AI-driven CI has already made significant impacts:
Medical Diagnostics : AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making.
: AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making. Misinformation Detection : In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time.
: In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time. Environmental Conservation: Projects such as Litterati and eBird leverage AI-enhanced CI to engage communities in identifying pollution patterns and monitoring biodiversity, respectively. AI analyzes large datasets generated by human contributors, leading to more effective conservation strategies.
These examples illustrate the potential of AI-enhanced CI to drive efficiency, accuracy, and scalability in diverse fields, offering innovative solutions to global challenges.
Challenges and future directions
Despite its promise, AI-enhanced CI faces several challenges. The study underscores concerns related to AI bias, transparency, and ethical considerations. AI models are only as good as the data they are trained on, and biases in data can lead to flawed decision-making. Additionally, the lack of explainability in AI-generated outcomes raises questions about accountability and trust in human-AI collaborations.
Another challenge is maintaining human motivation and engagement in AI-integrated decision-making systems. Over-reliance on AI could lead to diminished human participation, reducing the benefits of diverse perspectives and ethical reasoning that human collectives bring to the table.
The future of AI-enhanced CI lies in refining human-AI interaction frameworks to ensure optimal collaboration. Researchers call for the development of AI systems that are transparent, interpretable, and aligned with human values. Moreover, interdisciplinary cooperation between cognitive scientists, ethicists, and AI developers is necessary to design CI systems that balance computational efficiency with ethical and social considerations.
Conclusion: The path forward
The study by Cui and Yasseri presents a compelling vision for the future of collective intelligence, where AI is not a replacement for human intellect but a catalyst for enhanced decision-making. By fostering a symbiotic relationship between AI and human cognition, society can unlock new levels of problem-solving capabilities, addressing global challenges with greater efficiency and precision.
As AI continues to evolve, the focus must shift toward creating hybrid intelligence systems that integrate AI’s strengths while preserving human values, creativity, and ethical judgment. The ultimate goal is not to substitute human intelligence but to augment it, ensuring that the combined power of AI and human collectives drives progress in a rapidly changing world. | 5 |
In an age where complex societal challenges exceed human capabilities, the integration of artificial intelligence (AI) into collective decision-making has emerged as a transformative force. While some fear AI replacing human intellect, a groundbreaking study argues that AI, when integrated effectively, can enhance human collective intelligence rather than replace it.
The study, titled AI-enhanced Collective Intelligence, authored by Hao Cui and Taha Yasseri and published in Patterns, explores the synergy between human cognition and AI, advocating for a hybrid model where AI augments human capabilities to tackle increasingly intricate problems.
The evolution of collective intelligence
Collective intelligence (CI) has long been recognized as the superior outcome of collaborative human efforts. From ancient social structures to modern technological advancements, humanity has leveraged collective knowledge to drive progress. The Internet and large-scale collaboration platforms, such as Wikipedia and crowdsourced research projects, have already elevated CI by enabling knowledge sharing across vast networks.
However, the advent of AI introduces new dimensions to CI. Unlike traditional models, where human collectives function independently, AI-enhanced CI proposes an interconnected system where AI and human agents coalesce to form superior decision-making units. AI brings computational efficiency, pattern recognition, and data-processing prowess, while humans contribute intuition, creativity, and ethical considerations. Together, these complementary capabilities can yield unprecedented levels of intelligence and problem-solving efficacy.
Understanding the human-AI hybrid model
The study presents a multilayer representation of AI-enhanced CI, structured around three core layers: cognition, physical, and information. In this model, humans and AI interact dynamically within and across these layers, forming an interconnected system that amplifies collective problem-solving potential.
Cognition Layer: This includes the mental processes of both humans and AI, encompassing decision-making, sense-making, and reasoning. AI can enhance cognition by providing vast amounts of structured information, aiding in knowledge synthesis. Physical Layer: This represents tangible interactions between humans and AI, such as AI-driven robotic automation in industries or real-world collaboration in diagnostics and engineering. Information Layer: This layer governs the exchange of data between human and AI agents, determining how insights are shared and integrated to enhance CI.
By conceptualizing CI as a networked system, the study highlights how AI’s consistency, scalability, and adaptability can complement human intelligence, reducing biases and inconsistencies often present in human decision-making.
Applications and real-world impact
The real-world implications of AI-enhanced CI span multiple domains, from scientific research to policy-making, healthcare, and environmental sustainability. The study examines cases where AI-driven CI has already made significant impacts:
Medical Diagnostics : AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making.
: AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making. Misinformation Detection : In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time.
: In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time. Environmental Conservation: Projects such as Litterati and eBird leverage AI-enhanced CI to engage communities in identifying pollution patterns and monitoring biodiversity, respectively. AI analyzes large datasets generated by human contributors, leading to more effective conservation strategies.
These examples illustrate the potential of AI-enhanced CI to drive efficiency, accuracy, and scalability in diverse fields, offering innovative solutions to global challenges.
Challenges and future directions
Despite its promise, AI-enhanced CI faces several challenges. The study underscores concerns related to AI bias, transparency, and ethical considerations. AI models are only as good as the data they are trained on, and biases in data can lead to flawed decision-making. Additionally, the lack of explainability in AI-generated outcomes raises questions about accountability and trust in human-AI collaborations.
Another challenge is maintaining human motivation and engagement in AI-integrated decision-making systems. Over-reliance on AI could lead to diminished human participation, reducing the benefits of diverse perspectives and ethical reasoning that human collectives bring to the table.
The future of AI-enhanced CI lies in refining human-AI interaction frameworks to ensure optimal collaboration. Researchers call for the development of AI systems that are transparent, interpretable, and aligned with human values. Moreover, interdisciplinary cooperation between cognitive scientists, ethicists, and AI developers is necessary to design CI systems that balance computational efficiency with ethical and social considerations.
Conclusion: The path forward
The study by Cui and Yasseri presents a compelling vision for the future of collective intelligence, where AI is not a replacement for human intellect but a catalyst for enhanced decision-making. By fostering a symbiotic relationship between AI and human cognition, society can unlock new levels of problem-solving capabilities, addressing global challenges with greater efficiency and precision.
As AI continues to evolve, the focus must shift toward creating hybrid intelligence systems that integrate AI’s strengths while preserving human values, creativity, and ethical judgment. The ultimate goal is not to substitute human intelligence but to augment it, ensuring that the combined power of AI and human collectives drives progress in a rapidly changing world. | 5 |
In an age where complex societal challenges exceed human capabilities, the integration of artificial intelligence (AI) into collective decision-making has emerged as a transformative force. While some fear AI replacing human intellect, a groundbreaking study argues that AI, when integrated effectively, can enhance human collective intelligence rather than replace it.
The study, titled AI-enhanced Collective Intelligence, authored by Hao Cui and Taha Yasseri and published in Patterns, explores the synergy between human cognition and AI, advocating for a hybrid model where AI augments human capabilities to tackle increasingly intricate problems.
The evolution of collective intelligence
Collective intelligence (CI) has long been recognized as the superior outcome of collaborative human efforts. From ancient social structures to modern technological advancements, humanity has leveraged collective knowledge to drive progress. The Internet and large-scale collaboration platforms, such as Wikipedia and crowdsourced research projects, have already elevated CI by enabling knowledge sharing across vast networks.
However, the advent of AI introduces new dimensions to CI. Unlike traditional models, where human collectives function independently, AI-enhanced CI proposes an interconnected system where AI and human agents coalesce to form superior decision-making units. AI brings computational efficiency, pattern recognition, and data-processing prowess, while humans contribute intuition, creativity, and ethical considerations. Together, these complementary capabilities can yield unprecedented levels of intelligence and problem-solving efficacy.
Understanding the human-AI hybrid model
The study presents a multilayer representation of AI-enhanced CI, structured around three core layers: cognition, physical, and information. In this model, humans and AI interact dynamically within and across these layers, forming an interconnected system that amplifies collective problem-solving potential.
Cognition Layer: This includes the mental processes of both humans and AI, encompassing decision-making, sense-making, and reasoning. AI can enhance cognition by providing vast amounts of structured information, aiding in knowledge synthesis. Physical Layer: This represents tangible interactions between humans and AI, such as AI-driven robotic automation in industries or real-world collaboration in diagnostics and engineering. Information Layer: This layer governs the exchange of data between human and AI agents, determining how insights are shared and integrated to enhance CI.
By conceptualizing CI as a networked system, the study highlights how AI’s consistency, scalability, and adaptability can complement human intelligence, reducing biases and inconsistencies often present in human decision-making.
Applications and real-world impact
The real-world implications of AI-enhanced CI span multiple domains, from scientific research to policy-making, healthcare, and environmental sustainability. The study examines cases where AI-driven CI has already made significant impacts:
Medical Diagnostics : AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making.
: AI-driven collective intelligence is being used to aggregate insights from global medical experts, improving diagnostic accuracy and patient outcomes. Platforms like Human Diagnosis Project harness the expertise of doctors and AI algorithms to enhance medical decision-making. Misinformation Detection : In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time.
: In an era of rampant misinformation, AI-augmented fact-checking systems, such as those used by investigative journalism groups like Bellingcat, improve the accuracy of information dissemination by cross-referencing sources in real time. Environmental Conservation: Projects such as Litterati and eBird leverage AI-enhanced CI to engage communities in identifying pollution patterns and monitoring biodiversity, respectively. AI analyzes large datasets generated by human contributors, leading to more effective conservation strategies.
These examples illustrate the potential of AI-enhanced CI to drive efficiency, accuracy, and scalability in diverse fields, offering innovative solutions to global challenges.
Challenges and future directions
Despite its promise, AI-enhanced CI faces several challenges. The study underscores concerns related to AI bias, transparency, and ethical considerations. AI models are only as good as the data they are trained on, and biases in data can lead to flawed decision-making. Additionally, the lack of explainability in AI-generated outcomes raises questions about accountability and trust in human-AI collaborations.
Another challenge is maintaining human motivation and engagement in AI-integrated decision-making systems. Over-reliance on AI could lead to diminished human participation, reducing the benefits of diverse perspectives and ethical reasoning that human collectives bring to the table.
The future of AI-enhanced CI lies in refining human-AI interaction frameworks to ensure optimal collaboration. Researchers call for the development of AI systems that are transparent, interpretable, and aligned with human values. Moreover, interdisciplinary cooperation between cognitive scientists, ethicists, and AI developers is necessary to design CI systems that balance computational efficiency with ethical and social considerations.
Conclusion: The path forward
The study by Cui and Yasseri presents a compelling vision for the future of collective intelligence, where AI is not a replacement for human intellect but a catalyst for enhanced decision-making. By fostering a symbiotic relationship between AI and human cognition, society can unlock new levels of problem-solving capabilities, addressing global challenges with greater efficiency and precision.
As AI continues to evolve, the focus must shift toward creating hybrid intelligence systems that integrate AI’s strengths while preserving human values, creativity, and ethical judgment. The ultimate goal is not to substitute human intelligence but to augment it, ensuring that the combined power of AI and human collectives drives progress in a rapidly changing world. | 5 |
Amidst the global competition surrounding artificial intelligence, India’s ambitions in the domain have now become impossible to ignore. While India has thus far been a modest player in the global AI race, recent developments—particularly after the advent of DeepSeek AI—show that the country is making serious strides in the direction of putting itself at par with countries like the US and China. The IndiaAI Mission, with over USD 1.2 billion in investment, is pushing India to the forefront of AI innovation. Startups across the country are gaining global attention, and corporations—major and minor—are betting big on India’s AI future.
Experts are of the opinion that India is well-positioned to lead in AI, both as a participant and a key decision-maker. India’s role as co-chair at the recent Paris AI Summit is indicative of its growing influence. Unlike the US and the UK, where AI breakthroughs are often research-driven, India’s approach is significantly different. Indian companies are focused on large-scale adoption and cost-effective solutions, which will make AI accessible to more people. The country’s talent pool is also among the advantages. With the second-largest AI workforce globally, India has no shortage of skilled engineers and researchers. Cities like Bangalore, Hyderabad, and Pune are becoming AI hubs. While India may not yet be a leader in foundational AI research, it is making major contributions in applying AI across industries like healthcare, finance, and agriculture, among others. The recent launch of the IndiaAI Mission is a strong statement regarding the country’s commitment to build AI infrastructure. India’s massive, young, and tech-savvy population is driving AI adoption. More than half of the country’s citizens are under 30, potentially serving as a huge market for AI-powered services. Companies are using AI to transform everyday experiences, from banking to retail to transportation. AI-driven consumer technology is rapidly changing business models and redefining industries. One of India’s biggest strengths in AI is its linguistic diversity. Unlike global AI models that focus primarily on English, India has a unique opportunity to build AI in multiple languages. Local AI models can be tailored to regional needs. This will help preserve cultural diversity and ensure enhanced accessibility. Funding innovative models could help Indian firms create AI tools that are locally relevant and owned by Indian companies.
Needless to say, despite this progress, there are significant challenges ahead. Many AI projects require heavy investment in infrastructure. Competing with AI powerhouses like the US and China will take time and significant resources. Experts argue that more government-led R&D is necessary, as was done in the past successful projects like UPI and Aadhaar. AI development also comes with ethical and regulatory concerns. Balancing rapid growth with responsible AI governance will be crucial. India has an opportunity to set a new global blueprint for AI development. Public-private partnerships, investments from global tech giants, and a focus on frugal innovation give the country a strong foundation. The AI race is not just about who can build the biggest models but about who can create meaningful, real-world solutions. With its vast talent pool, entrepreneurial spirit, and commitment to affordable, practical AI, India can position itself as a leader in shaping the future of artificial intelligence. | 5 |
Amidst the global competition surrounding artificial intelligence, India’s ambitions in the domain have now become impossible to ignore. While India has thus far been a modest player in the global AI race, recent developments—particularly after the advent of DeepSeek AI—show that the country is making serious strides in the direction of putting itself at par with countries like the US and China. The IndiaAI Mission, with over USD 1.2 billion in investment, is pushing India to the forefront of AI innovation. Startups across the country are gaining global attention, and corporations—major and minor—are betting big on India’s AI future.
Experts are of the opinion that India is well-positioned to lead in AI, both as a participant and a key decision-maker. India’s role as co-chair at the recent Paris AI Summit is indicative of its growing influence. Unlike the US and the UK, where AI breakthroughs are often research-driven, India’s approach is significantly different. Indian companies are focused on large-scale adoption and cost-effective solutions, which will make AI accessible to more people. The country’s talent pool is also among the advantages. With the second-largest AI workforce globally, India has no shortage of skilled engineers and researchers. Cities like Bangalore, Hyderabad, and Pune are becoming AI hubs. While India may not yet be a leader in foundational AI research, it is making major contributions in applying AI across industries like healthcare, finance, and agriculture, among others. The recent launch of the IndiaAI Mission is a strong statement regarding the country’s commitment to build AI infrastructure. India’s massive, young, and tech-savvy population is driving AI adoption. More than half of the country’s citizens are under 30, potentially serving as a huge market for AI-powered services. Companies are using AI to transform everyday experiences, from banking to retail to transportation. AI-driven consumer technology is rapidly changing business models and redefining industries. One of India’s biggest strengths in AI is its linguistic diversity. Unlike global AI models that focus primarily on English, India has a unique opportunity to build AI in multiple languages. Local AI models can be tailored to regional needs. This will help preserve cultural diversity and ensure enhanced accessibility. Funding innovative models could help Indian firms create AI tools that are locally relevant and owned by Indian companies.
Needless to say, despite this progress, there are significant challenges ahead. Many AI projects require heavy investment in infrastructure. Competing with AI powerhouses like the US and China will take time and significant resources. Experts argue that more government-led R&D is necessary, as was done in the past successful projects like UPI and Aadhaar. AI development also comes with ethical and regulatory concerns. Balancing rapid growth with responsible AI governance will be crucial. India has an opportunity to set a new global blueprint for AI development. Public-private partnerships, investments from global tech giants, and a focus on frugal innovation give the country a strong foundation. The AI race is not just about who can build the biggest models but about who can create meaningful, real-world solutions. With its vast talent pool, entrepreneurial spirit, and commitment to affordable, practical AI, India can position itself as a leader in shaping the future of artificial intelligence. | 5 |
Amidst the global competition surrounding artificial intelligence, India’s ambitions in the domain have now become impossible to ignore. While India has thus far been a modest player in the global AI race, recent developments—particularly after the advent of DeepSeek AI—show that the country is making serious strides in the direction of putting itself at par with countries like the US and China. The IndiaAI Mission, with over USD 1.2 billion in investment, is pushing India to the forefront of AI innovation. Startups across the country are gaining global attention, and corporations—major and minor—are betting big on India’s AI future.
Experts are of the opinion that India is well-positioned to lead in AI, both as a participant and a key decision-maker. India’s role as co-chair at the recent Paris AI Summit is indicative of its growing influence. Unlike the US and the UK, where AI breakthroughs are often research-driven, India’s approach is significantly different. Indian companies are focused on large-scale adoption and cost-effective solutions, which will make AI accessible to more people. The country’s talent pool is also among the advantages. With the second-largest AI workforce globally, India has no shortage of skilled engineers and researchers. Cities like Bangalore, Hyderabad, and Pune are becoming AI hubs. While India may not yet be a leader in foundational AI research, it is making major contributions in applying AI across industries like healthcare, finance, and agriculture, among others. The recent launch of the IndiaAI Mission is a strong statement regarding the country’s commitment to build AI infrastructure. India’s massive, young, and tech-savvy population is driving AI adoption. More than half of the country’s citizens are under 30, potentially serving as a huge market for AI-powered services. Companies are using AI to transform everyday experiences, from banking to retail to transportation. AI-driven consumer technology is rapidly changing business models and redefining industries. One of India’s biggest strengths in AI is its linguistic diversity. Unlike global AI models that focus primarily on English, India has a unique opportunity to build AI in multiple languages. Local AI models can be tailored to regional needs. This will help preserve cultural diversity and ensure enhanced accessibility. Funding innovative models could help Indian firms create AI tools that are locally relevant and owned by Indian companies.
Needless to say, despite this progress, there are significant challenges ahead. Many AI projects require heavy investment in infrastructure. Competing with AI powerhouses like the US and China will take time and significant resources. Experts argue that more government-led R&D is necessary, as was done in the past successful projects like UPI and Aadhaar. AI development also comes with ethical and regulatory concerns. Balancing rapid growth with responsible AI governance will be crucial. India has an opportunity to set a new global blueprint for AI development. Public-private partnerships, investments from global tech giants, and a focus on frugal innovation give the country a strong foundation. The AI race is not just about who can build the biggest models but about who can create meaningful, real-world solutions. With its vast talent pool, entrepreneurial spirit, and commitment to affordable, practical AI, India can position itself as a leader in shaping the future of artificial intelligence. | 5 |
Amidst the global competition surrounding artificial intelligence, India’s ambitions in the domain have now become impossible to ignore. While India has thus far been a modest player in the global AI race, recent developments—particularly after the advent of DeepSeek AI—show that the country is making serious strides in the direction of putting itself at par with countries like the US and China. The IndiaAI Mission, with over USD 1.2 billion in investment, is pushing India to the forefront of AI innovation. Startups across the country are gaining global attention, and corporations—major and minor—are betting big on India’s AI future.
Experts are of the opinion that India is well-positioned to lead in AI, both as a participant and a key decision-maker. India’s role as co-chair at the recent Paris AI Summit is indicative of its growing influence. Unlike the US and the UK, where AI breakthroughs are often research-driven, India’s approach is significantly different. Indian companies are focused on large-scale adoption and cost-effective solutions, which will make AI accessible to more people. The country’s talent pool is also among the advantages. With the second-largest AI workforce globally, India has no shortage of skilled engineers and researchers. Cities like Bangalore, Hyderabad, and Pune are becoming AI hubs. While India may not yet be a leader in foundational AI research, it is making major contributions in applying AI across industries like healthcare, finance, and agriculture, among others. The recent launch of the IndiaAI Mission is a strong statement regarding the country’s commitment to build AI infrastructure. India’s massive, young, and tech-savvy population is driving AI adoption. More than half of the country’s citizens are under 30, potentially serving as a huge market for AI-powered services. Companies are using AI to transform everyday experiences, from banking to retail to transportation. AI-driven consumer technology is rapidly changing business models and redefining industries. One of India’s biggest strengths in AI is its linguistic diversity. Unlike global AI models that focus primarily on English, India has a unique opportunity to build AI in multiple languages. Local AI models can be tailored to regional needs. This will help preserve cultural diversity and ensure enhanced accessibility. Funding innovative models could help Indian firms create AI tools that are locally relevant and owned by Indian companies.
Needless to say, despite this progress, there are significant challenges ahead. Many AI projects require heavy investment in infrastructure. Competing with AI powerhouses like the US and China will take time and significant resources. Experts argue that more government-led R&D is necessary, as was done in the past successful projects like UPI and Aadhaar. AI development also comes with ethical and regulatory concerns. Balancing rapid growth with responsible AI governance will be crucial. India has an opportunity to set a new global blueprint for AI development. Public-private partnerships, investments from global tech giants, and a focus on frugal innovation give the country a strong foundation. The AI race is not just about who can build the biggest models but about who can create meaningful, real-world solutions. With its vast talent pool, entrepreneurial spirit, and commitment to affordable, practical AI, India can position itself as a leader in shaping the future of artificial intelligence. | 5 |
Amidst the global competition surrounding artificial intelligence, India’s ambitions in the domain have now become impossible to ignore. While India has thus far been a modest player in the global AI race, recent developments—particularly after the advent of DeepSeek AI—show that the country is making serious strides in the direction of putting itself at par with countries like the US and China. The IndiaAI Mission, with over USD 1.2 billion in investment, is pushing India to the forefront of AI innovation. Startups across the country are gaining global attention, and corporations—major and minor—are betting big on India’s AI future.
Experts are of the opinion that India is well-positioned to lead in AI, both as a participant and a key decision-maker. India’s role as co-chair at the recent Paris AI Summit is indicative of its growing influence. Unlike the US and the UK, where AI breakthroughs are often research-driven, India’s approach is significantly different. Indian companies are focused on large-scale adoption and cost-effective solutions, which will make AI accessible to more people. The country’s talent pool is also among the advantages. With the second-largest AI workforce globally, India has no shortage of skilled engineers and researchers. Cities like Bangalore, Hyderabad, and Pune are becoming AI hubs. While India may not yet be a leader in foundational AI research, it is making major contributions in applying AI across industries like healthcare, finance, and agriculture, among others. The recent launch of the IndiaAI Mission is a strong statement regarding the country’s commitment to build AI infrastructure. India’s massive, young, and tech-savvy population is driving AI adoption. More than half of the country’s citizens are under 30, potentially serving as a huge market for AI-powered services. Companies are using AI to transform everyday experiences, from banking to retail to transportation. AI-driven consumer technology is rapidly changing business models and redefining industries. One of India’s biggest strengths in AI is its linguistic diversity. Unlike global AI models that focus primarily on English, India has a unique opportunity to build AI in multiple languages. Local AI models can be tailored to regional needs. This will help preserve cultural diversity and ensure enhanced accessibility. Funding innovative models could help Indian firms create AI tools that are locally relevant and owned by Indian companies.
Needless to say, despite this progress, there are significant challenges ahead. Many AI projects require heavy investment in infrastructure. Competing with AI powerhouses like the US and China will take time and significant resources. Experts argue that more government-led R&D is necessary, as was done in the past successful projects like UPI and Aadhaar. AI development also comes with ethical and regulatory concerns. Balancing rapid growth with responsible AI governance will be crucial. India has an opportunity to set a new global blueprint for AI development. Public-private partnerships, investments from global tech giants, and a focus on frugal innovation give the country a strong foundation. The AI race is not just about who can build the biggest models but about who can create meaningful, real-world solutions. With its vast talent pool, entrepreneurial spirit, and commitment to affordable, practical AI, India can position itself as a leader in shaping the future of artificial intelligence. | 5 |
Artificial Intelligence (AI) is the biggest technological shift we have seen in our lifetime to date and will be the driving force behind India's economic growth, said Akash Ambani, Chairman of Reliance Jio Infocomm, at the Mumbai Tech Week held at Jio World Centre.
Also Read: Mumbai Set to Host Asia’s Largest AI Conference – MTW 2025
Investing in AI, Research, and Talent
"In my view, AI is the engine that will empower India to grow at 10 percent or double-digit growth numbers for the foreseeable future," Ambani stated, noting that "India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
To establish India as a global AI leader, Ambani outlined three key areas of focus: infrastructure, deep research and development, and talent investment.
Akash Ambani, Chairman of Reliance Jio Infocomm, spoke during a fireside chat with Dream11 CEO Harsh Jain at the Mumbai Tech Week held at Jio World Centre.
Jio’s Impact on India's Digital Transformation
Opening the conversation about Jio and its achievements in India, Dream11 CEO Harsh Jain said that Jio is literally at the forefront of technology in India. "Every single person here is thankful for finally getting us 5G Internet at the best speeds possible while we travel. I get like 150 Mbps in my car, on my phone, which is just phenomenal," Jain said.
"..We're all very fortunate to be in a country where our domestic product itself can reach 800 million Indians thanks to Jio. And it's a great time to be running a tech company and to have tech startups for all the entrepreneurs out there," Jain added.
India is a Leader in Adopting Technology
Responding to Jain's question about India still being seen as a laggard in technology, Akash Ambani replied, "I feel the contrary."
"Today, eight years after launching Jio, we've become the biggest data-consuming nation in the world. We've led that every, every single hour, average consumption per user is now significantly higher than anywhere else in the world, including China. So I think gone are the days that we should think about India from a tech laggard perspective. I think we have established that India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
Also Read: Reliance AI Infrastructure in Jamnagar to Be Built in 24 Months: Akash Ambani
Ambani further stated that as a fundamental block, "We need to continue investing in AI infrastructure and AI data centers that are fully equipped to scale globally and support millions of users in India. At Jio, we are already doing this. We recently announced in Jamnagar that we are building our AI data center, which will have a gigawatt capacity."
He also stressed the importance of deep research and development. "We continue to invest in deep research and deep development that comes from it," he said, crediting Prime Minister Narendra Modi's vision for AI advancement through the AI mission.
Additionally, Ambani highlighted the need to invest in the right talent. "At Jio, we've already invested in our overall full-stack AI team, which is led by data scientists, researchers and engineers to be a thousand plus. The critical element in this, I feel, is also to foster new ideas and push the boundaries of development," he said, adding, "It's not too far away where we will have a groundbreaking idea that will have half a billion people on one platform coming out of India."
Perception of India as a Tech Laggard
Challenging the perception that India lags in technology adoption, Ambani pointed to the country's rapid digital transformation. "I think we've already showcased to the world in connectivity that we can be the leaders of technology, not just be fast followers."
He recalled India's shift in internet usage and speeds since 2015, saying, "...Back to 2015, which is just 10 years ago, Internet speeds were less than 1 MB in this country, whether it be on the mobile or at home. But today, everyone continues to enjoy very high data speeds. And so I think from an AI point of view is these three critical items that we need to focus on," Ambani said.
Replying to Ambani, Jain said, "But then when we grew up and we saw speeds in America, we'd be like, what the hell? Why is our speed, you know, why is our Internet speed like this in India? And why is this like, apparently 3G giving me less than 1 Mbps? And today I can say that when we travel, we laugh at the world. We literally were there and we are like, wait, you pay 60-70 USD a month for crappy Internet? Boss, come to India, just try Jio for a while and then you'll see what we have. So thank you for that. It's been an amazing thing to see all of India benefit from this."
Also Read: Akash Ambani Urges Rapid Adoption of AI and Data Centre Policy Reform in India
AI Taking Over Jobs
Responding to Jain's question about whether AI will eliminate many jobs, Ambani said, "I'm a firm believer that AI will transform jobs. Today we've seen AI take over our mundane tasks, our repetitive tasks."
"We at Jio are already embracing it," Ambani said, adding, "We've seen how the Internet can create new industries and jobs, but this time, we can prepare for it. And by preparing, I mean equipping ourselves with the right tools and insights to excel our impact, excel our businesses. For example, at Jio, today we monitor our network, which now covers about 95 percent of the living population of India."
Leveraging AI for Network Monitoring
Before that, we had to monitor networks across different systems. "Today we can proactively monitor and before even a customer has a bad experience, experience, we can predict it. And this is just not, you know, something that is, this is again, instead of deeping/delving deep into the data, figuring out what the consumer does, the patterns are being picked up by ML and AI engineers to kind of spend the time to actually solve the problem than finding out what the problem is. So that's one way that we prepared," Ambani explained.
AI in Education
Responding to Jain's idea of leveraging AI to solve education challenges in India, Ambani said, "For us, there are five layers that create a deep technology company. In this particular use case, the first thing to solve for in education is connectivity, which we pretty much solved in India."
Five Layers
Today, 1.5 million schools have connectivity. The second layer is computing—both data center computing and edge computing. At Jio, we are advancing in both across the nation. "After that comes the devices layer where we have to enable devices that can consume and actually deliver to the endpoint which in this case will be the student across, affordability, standpoints across," Ambani explained, referring to low-cost consumer devices.
"So it may be a cloud PC laptop or a low-cost laptop. It could be a smartphone tomorrow. It can be AR glasses that we can use to deliver these devices across to enhance the experience that
we can actually deliver. On top of that comes the content layer where content not in the form of videos but just educational content or anything. And then comes the intelligence layer. So these five layers put together, you know, enable us to become solving big problems like education," Ambani said.
"And finally the five layers that I spoke about, they need to create an absolutely seamless customer experience. If they do not create a seamless customer experience that is easy to adopt, easy to have, then it becomes very, very hard for us to solve large issues like the literacy rate and education in India. We're going to be doing that," he added.
He further said the same approach applies to agriculture, IoT devices, and healthcare. At Jio, we will work to scale solutions and impact millions of students and schools in India, Ambani said.
Aspiring Indian: AI
Jain expressed that the future of AI and technology in India will only be realized through full collaboration between public and private sectors.
Ambani responded, "We're extremely lucky to be under the visionary leadership of our Prime minister Narendra Modi on this." I think, it is the biggest blessing of our country to have a leader like him leading this mission. "Recently, at Parliament, he mentioned that AI does not only just start for artificial intelligence. It actually stands for Aspiring Indian."
Jio as a Platform Company
When asked whether Jio aims to become a platform company like Google, Amazon, Apple, or Meta, Ambani said, "At Jio, we also imagine ourselves to be platform companies. One of our biggest growth missions
of Jio and the way where we want to make impact is connecting each home. And in the home, we not only offer broadband connectivity but we offer our own OS which is a Tele OS that powers the large screen at the home."
"It is our vision to reach 100 million homes in the shortest amount of time possible. And I think we have now the technology and the demand to do so to enable that."
Replying to Ambani, Jain said, Take UPI, for example. We've almost started to take it for granted. But when you travel, you realize how far behind the world is in peer-to-peer payments. India has leapfrogged ahead.
"The way Jio took us from, you know, 3G, which wasn't working properly, to like 5G, which is the world's best…in a few years we leapfrogged that entire technology wave. And I can tell you all it's a small secret between 5,000 of us. But you know, this is not just a vision, this is actually something that, that Akash has running in his home. What you're talking, what he's talking about here is like, you know, have any of you all seen Iron man movie with Jarvis. Right? That's the kind of stuff they already have and they're working on to bring to our homes," Jain said, addressing the audience.
Also Read: Jio Platforms to Soon Launch Cloud-Based AI Personal Computer: Akash Ambani
Cloud-based AI PC
Ambani then spoke about enabling entrepreneurship. One possibility is GPU as a service—creating a foundation for developers to build on. Similarly, we will soon launch a cloud-based AI PC. You can read more about the Cloud-Based AI PC in the story linked above.
Jio Campus Open for Visits
He also announced that Jio's Mumbai campus, spanning 500 acres and home to 25,000 people, including 10,000–12,000 Jio employees, is now open for public visits. | 5 |
Artificial Intelligence (AI) is the biggest technological shift we have seen in our lifetime to date and will be the driving force behind India's economic growth, said Akash Ambani, Chairman of Reliance Jio Infocomm, at the Mumbai Tech Week held at Jio World Centre.
Also Read: Mumbai Set to Host Asia’s Largest AI Conference – MTW 2025
Investing in AI, Research, and Talent
"In my view, AI is the engine that will empower India to grow at 10 percent or double-digit growth numbers for the foreseeable future," Ambani stated, noting that "India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
To establish India as a global AI leader, Ambani outlined three key areas of focus: infrastructure, deep research and development, and talent investment.
Akash Ambani, Chairman of Reliance Jio Infocomm, spoke during a fireside chat with Dream11 CEO Harsh Jain at the Mumbai Tech Week held at Jio World Centre.
Jio’s Impact on India's Digital Transformation
Opening the conversation about Jio and its achievements in India, Dream11 CEO Harsh Jain said that Jio is literally at the forefront of technology in India. "Every single person here is thankful for finally getting us 5G Internet at the best speeds possible while we travel. I get like 150 Mbps in my car, on my phone, which is just phenomenal," Jain said.
"..We're all very fortunate to be in a country where our domestic product itself can reach 800 million Indians thanks to Jio. And it's a great time to be running a tech company and to have tech startups for all the entrepreneurs out there," Jain added.
India is a Leader in Adopting Technology
Responding to Jain's question about India still being seen as a laggard in technology, Akash Ambani replied, "I feel the contrary."
"Today, eight years after launching Jio, we've become the biggest data-consuming nation in the world. We've led that every, every single hour, average consumption per user is now significantly higher than anywhere else in the world, including China. So I think gone are the days that we should think about India from a tech laggard perspective. I think we have established that India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
Also Read: Reliance AI Infrastructure in Jamnagar to Be Built in 24 Months: Akash Ambani
Ambani further stated that as a fundamental block, "We need to continue investing in AI infrastructure and AI data centers that are fully equipped to scale globally and support millions of users in India. At Jio, we are already doing this. We recently announced in Jamnagar that we are building our AI data center, which will have a gigawatt capacity."
He also stressed the importance of deep research and development. "We continue to invest in deep research and deep development that comes from it," he said, crediting Prime Minister Narendra Modi's vision for AI advancement through the AI mission.
Additionally, Ambani highlighted the need to invest in the right talent. "At Jio, we've already invested in our overall full-stack AI team, which is led by data scientists, researchers and engineers to be a thousand plus. The critical element in this, I feel, is also to foster new ideas and push the boundaries of development," he said, adding, "It's not too far away where we will have a groundbreaking idea that will have half a billion people on one platform coming out of India."
Perception of India as a Tech Laggard
Challenging the perception that India lags in technology adoption, Ambani pointed to the country's rapid digital transformation. "I think we've already showcased to the world in connectivity that we can be the leaders of technology, not just be fast followers."
He recalled India's shift in internet usage and speeds since 2015, saying, "...Back to 2015, which is just 10 years ago, Internet speeds were less than 1 MB in this country, whether it be on the mobile or at home. But today, everyone continues to enjoy very high data speeds. And so I think from an AI point of view is these three critical items that we need to focus on," Ambani said.
Replying to Ambani, Jain said, "But then when we grew up and we saw speeds in America, we'd be like, what the hell? Why is our speed, you know, why is our Internet speed like this in India? And why is this like, apparently 3G giving me less than 1 Mbps? And today I can say that when we travel, we laugh at the world. We literally were there and we are like, wait, you pay 60-70 USD a month for crappy Internet? Boss, come to India, just try Jio for a while and then you'll see what we have. So thank you for that. It's been an amazing thing to see all of India benefit from this."
Also Read: Akash Ambani Urges Rapid Adoption of AI and Data Centre Policy Reform in India
AI Taking Over Jobs
Responding to Jain's question about whether AI will eliminate many jobs, Ambani said, "I'm a firm believer that AI will transform jobs. Today we've seen AI take over our mundane tasks, our repetitive tasks."
"We at Jio are already embracing it," Ambani said, adding, "We've seen how the Internet can create new industries and jobs, but this time, we can prepare for it. And by preparing, I mean equipping ourselves with the right tools and insights to excel our impact, excel our businesses. For example, at Jio, today we monitor our network, which now covers about 95 percent of the living population of India."
Leveraging AI for Network Monitoring
Before that, we had to monitor networks across different systems. "Today we can proactively monitor and before even a customer has a bad experience, experience, we can predict it. And this is just not, you know, something that is, this is again, instead of deeping/delving deep into the data, figuring out what the consumer does, the patterns are being picked up by ML and AI engineers to kind of spend the time to actually solve the problem than finding out what the problem is. So that's one way that we prepared," Ambani explained.
AI in Education
Responding to Jain's idea of leveraging AI to solve education challenges in India, Ambani said, "For us, there are five layers that create a deep technology company. In this particular use case, the first thing to solve for in education is connectivity, which we pretty much solved in India."
Five Layers
Today, 1.5 million schools have connectivity. The second layer is computing—both data center computing and edge computing. At Jio, we are advancing in both across the nation. "After that comes the devices layer where we have to enable devices that can consume and actually deliver to the endpoint which in this case will be the student across, affordability, standpoints across," Ambani explained, referring to low-cost consumer devices.
"So it may be a cloud PC laptop or a low-cost laptop. It could be a smartphone tomorrow. It can be AR glasses that we can use to deliver these devices across to enhance the experience that
we can actually deliver. On top of that comes the content layer where content not in the form of videos but just educational content or anything. And then comes the intelligence layer. So these five layers put together, you know, enable us to become solving big problems like education," Ambani said.
"And finally the five layers that I spoke about, they need to create an absolutely seamless customer experience. If they do not create a seamless customer experience that is easy to adopt, easy to have, then it becomes very, very hard for us to solve large issues like the literacy rate and education in India. We're going to be doing that," he added.
He further said the same approach applies to agriculture, IoT devices, and healthcare. At Jio, we will work to scale solutions and impact millions of students and schools in India, Ambani said.
Aspiring Indian: AI
Jain expressed that the future of AI and technology in India will only be realized through full collaboration between public and private sectors.
Ambani responded, "We're extremely lucky to be under the visionary leadership of our Prime minister Narendra Modi on this." I think, it is the biggest blessing of our country to have a leader like him leading this mission. "Recently, at Parliament, he mentioned that AI does not only just start for artificial intelligence. It actually stands for Aspiring Indian."
Jio as a Platform Company
When asked whether Jio aims to become a platform company like Google, Amazon, Apple, or Meta, Ambani said, "At Jio, we also imagine ourselves to be platform companies. One of our biggest growth missions
of Jio and the way where we want to make impact is connecting each home. And in the home, we not only offer broadband connectivity but we offer our own OS which is a Tele OS that powers the large screen at the home."
"It is our vision to reach 100 million homes in the shortest amount of time possible. And I think we have now the technology and the demand to do so to enable that."
Replying to Ambani, Jain said, Take UPI, for example. We've almost started to take it for granted. But when you travel, you realize how far behind the world is in peer-to-peer payments. India has leapfrogged ahead.
"The way Jio took us from, you know, 3G, which wasn't working properly, to like 5G, which is the world's best…in a few years we leapfrogged that entire technology wave. And I can tell you all it's a small secret between 5,000 of us. But you know, this is not just a vision, this is actually something that, that Akash has running in his home. What you're talking, what he's talking about here is like, you know, have any of you all seen Iron man movie with Jarvis. Right? That's the kind of stuff they already have and they're working on to bring to our homes," Jain said, addressing the audience.
Also Read: Jio Platforms to Soon Launch Cloud-Based AI Personal Computer: Akash Ambani
Cloud-based AI PC
Ambani then spoke about enabling entrepreneurship. One possibility is GPU as a service—creating a foundation for developers to build on. Similarly, we will soon launch a cloud-based AI PC. You can read more about the Cloud-Based AI PC in the story linked above.
Jio Campus Open for Visits
He also announced that Jio's Mumbai campus, spanning 500 acres and home to 25,000 people, including 10,000–12,000 Jio employees, is now open for public visits. | 5 |
Artificial Intelligence (AI) is the biggest technological shift we have seen in our lifetime to date and will be the driving force behind India's economic growth, said Akash Ambani, Chairman of Reliance Jio Infocomm, at the Mumbai Tech Week held at Jio World Centre.
Also Read: Mumbai Set to Host Asia’s Largest AI Conference – MTW 2025
Investing in AI, Research, and Talent
"In my view, AI is the engine that will empower India to grow at 10 percent or double-digit growth numbers for the foreseeable future," Ambani stated, noting that "India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
To establish India as a global AI leader, Ambani outlined three key areas of focus: infrastructure, deep research and development, and talent investment.
Akash Ambani, Chairman of Reliance Jio Infocomm, spoke during a fireside chat with Dream11 CEO Harsh Jain at the Mumbai Tech Week held at Jio World Centre.
Jio’s Impact on India's Digital Transformation
Opening the conversation about Jio and its achievements in India, Dream11 CEO Harsh Jain said that Jio is literally at the forefront of technology in India. "Every single person here is thankful for finally getting us 5G Internet at the best speeds possible while we travel. I get like 150 Mbps in my car, on my phone, which is just phenomenal," Jain said.
"..We're all very fortunate to be in a country where our domestic product itself can reach 800 million Indians thanks to Jio. And it's a great time to be running a tech company and to have tech startups for all the entrepreneurs out there," Jain added.
India is a Leader in Adopting Technology
Responding to Jain's question about India still being seen as a laggard in technology, Akash Ambani replied, "I feel the contrary."
"Today, eight years after launching Jio, we've become the biggest data-consuming nation in the world. We've led that every, every single hour, average consumption per user is now significantly higher than anywhere else in the world, including China. So I think gone are the days that we should think about India from a tech laggard perspective. I think we have established that India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
Also Read: Reliance AI Infrastructure in Jamnagar to Be Built in 24 Months: Akash Ambani
Ambani further stated that as a fundamental block, "We need to continue investing in AI infrastructure and AI data centers that are fully equipped to scale globally and support millions of users in India. At Jio, we are already doing this. We recently announced in Jamnagar that we are building our AI data center, which will have a gigawatt capacity."
He also stressed the importance of deep research and development. "We continue to invest in deep research and deep development that comes from it," he said, crediting Prime Minister Narendra Modi's vision for AI advancement through the AI mission.
Additionally, Ambani highlighted the need to invest in the right talent. "At Jio, we've already invested in our overall full-stack AI team, which is led by data scientists, researchers and engineers to be a thousand plus. The critical element in this, I feel, is also to foster new ideas and push the boundaries of development," he said, adding, "It's not too far away where we will have a groundbreaking idea that will have half a billion people on one platform coming out of India."
Perception of India as a Tech Laggard
Challenging the perception that India lags in technology adoption, Ambani pointed to the country's rapid digital transformation. "I think we've already showcased to the world in connectivity that we can be the leaders of technology, not just be fast followers."
He recalled India's shift in internet usage and speeds since 2015, saying, "...Back to 2015, which is just 10 years ago, Internet speeds were less than 1 MB in this country, whether it be on the mobile or at home. But today, everyone continues to enjoy very high data speeds. And so I think from an AI point of view is these three critical items that we need to focus on," Ambani said.
Replying to Ambani, Jain said, "But then when we grew up and we saw speeds in America, we'd be like, what the hell? Why is our speed, you know, why is our Internet speed like this in India? And why is this like, apparently 3G giving me less than 1 Mbps? And today I can say that when we travel, we laugh at the world. We literally were there and we are like, wait, you pay 60-70 USD a month for crappy Internet? Boss, come to India, just try Jio for a while and then you'll see what we have. So thank you for that. It's been an amazing thing to see all of India benefit from this."
Also Read: Akash Ambani Urges Rapid Adoption of AI and Data Centre Policy Reform in India
AI Taking Over Jobs
Responding to Jain's question about whether AI will eliminate many jobs, Ambani said, "I'm a firm believer that AI will transform jobs. Today we've seen AI take over our mundane tasks, our repetitive tasks."
"We at Jio are already embracing it," Ambani said, adding, "We've seen how the Internet can create new industries and jobs, but this time, we can prepare for it. And by preparing, I mean equipping ourselves with the right tools and insights to excel our impact, excel our businesses. For example, at Jio, today we monitor our network, which now covers about 95 percent of the living population of India."
Leveraging AI for Network Monitoring
Before that, we had to monitor networks across different systems. "Today we can proactively monitor and before even a customer has a bad experience, experience, we can predict it. And this is just not, you know, something that is, this is again, instead of deeping/delving deep into the data, figuring out what the consumer does, the patterns are being picked up by ML and AI engineers to kind of spend the time to actually solve the problem than finding out what the problem is. So that's one way that we prepared," Ambani explained.
AI in Education
Responding to Jain's idea of leveraging AI to solve education challenges in India, Ambani said, "For us, there are five layers that create a deep technology company. In this particular use case, the first thing to solve for in education is connectivity, which we pretty much solved in India."
Five Layers
Today, 1.5 million schools have connectivity. The second layer is computing—both data center computing and edge computing. At Jio, we are advancing in both across the nation. "After that comes the devices layer where we have to enable devices that can consume and actually deliver to the endpoint which in this case will be the student across, affordability, standpoints across," Ambani explained, referring to low-cost consumer devices.
"So it may be a cloud PC laptop or a low-cost laptop. It could be a smartphone tomorrow. It can be AR glasses that we can use to deliver these devices across to enhance the experience that
we can actually deliver. On top of that comes the content layer where content not in the form of videos but just educational content or anything. And then comes the intelligence layer. So these five layers put together, you know, enable us to become solving big problems like education," Ambani said.
"And finally the five layers that I spoke about, they need to create an absolutely seamless customer experience. If they do not create a seamless customer experience that is easy to adopt, easy to have, then it becomes very, very hard for us to solve large issues like the literacy rate and education in India. We're going to be doing that," he added.
He further said the same approach applies to agriculture, IoT devices, and healthcare. At Jio, we will work to scale solutions and impact millions of students and schools in India, Ambani said.
Aspiring Indian: AI
Jain expressed that the future of AI and technology in India will only be realized through full collaboration between public and private sectors.
Ambani responded, "We're extremely lucky to be under the visionary leadership of our Prime minister Narendra Modi on this." I think, it is the biggest blessing of our country to have a leader like him leading this mission. "Recently, at Parliament, he mentioned that AI does not only just start for artificial intelligence. It actually stands for Aspiring Indian."
Jio as a Platform Company
When asked whether Jio aims to become a platform company like Google, Amazon, Apple, or Meta, Ambani said, "At Jio, we also imagine ourselves to be platform companies. One of our biggest growth missions
of Jio and the way where we want to make impact is connecting each home. And in the home, we not only offer broadband connectivity but we offer our own OS which is a Tele OS that powers the large screen at the home."
"It is our vision to reach 100 million homes in the shortest amount of time possible. And I think we have now the technology and the demand to do so to enable that."
Replying to Ambani, Jain said, Take UPI, for example. We've almost started to take it for granted. But when you travel, you realize how far behind the world is in peer-to-peer payments. India has leapfrogged ahead.
"The way Jio took us from, you know, 3G, which wasn't working properly, to like 5G, which is the world's best…in a few years we leapfrogged that entire technology wave. And I can tell you all it's a small secret between 5,000 of us. But you know, this is not just a vision, this is actually something that, that Akash has running in his home. What you're talking, what he's talking about here is like, you know, have any of you all seen Iron man movie with Jarvis. Right? That's the kind of stuff they already have and they're working on to bring to our homes," Jain said, addressing the audience.
Also Read: Jio Platforms to Soon Launch Cloud-Based AI Personal Computer: Akash Ambani
Cloud-based AI PC
Ambani then spoke about enabling entrepreneurship. One possibility is GPU as a service—creating a foundation for developers to build on. Similarly, we will soon launch a cloud-based AI PC. You can read more about the Cloud-Based AI PC in the story linked above.
Jio Campus Open for Visits
He also announced that Jio's Mumbai campus, spanning 500 acres and home to 25,000 people, including 10,000–12,000 Jio employees, is now open for public visits. | 5 |
Artificial Intelligence (AI) is the biggest technological shift we have seen in our lifetime to date and will be the driving force behind India's economic growth, said Akash Ambani, Chairman of Reliance Jio Infocomm, at the Mumbai Tech Week held at Jio World Centre.
Also Read: Mumbai Set to Host Asia’s Largest AI Conference – MTW 2025
Investing in AI, Research, and Talent
"In my view, AI is the engine that will empower India to grow at 10 percent or double-digit growth numbers for the foreseeable future," Ambani stated, noting that "India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
To establish India as a global AI leader, Ambani outlined three key areas of focus: infrastructure, deep research and development, and talent investment.
Akash Ambani, Chairman of Reliance Jio Infocomm, spoke during a fireside chat with Dream11 CEO Harsh Jain at the Mumbai Tech Week held at Jio World Centre.
Jio’s Impact on India's Digital Transformation
Opening the conversation about Jio and its achievements in India, Dream11 CEO Harsh Jain said that Jio is literally at the forefront of technology in India. "Every single person here is thankful for finally getting us 5G Internet at the best speeds possible while we travel. I get like 150 Mbps in my car, on my phone, which is just phenomenal," Jain said.
"..We're all very fortunate to be in a country where our domestic product itself can reach 800 million Indians thanks to Jio. And it's a great time to be running a tech company and to have tech startups for all the entrepreneurs out there," Jain added.
India is a Leader in Adopting Technology
Responding to Jain's question about India still being seen as a laggard in technology, Akash Ambani replied, "I feel the contrary."
"Today, eight years after launching Jio, we've become the biggest data-consuming nation in the world. We've led that every, every single hour, average consumption per user is now significantly higher than anywhere else in the world, including China. So I think gone are the days that we should think about India from a tech laggard perspective. I think we have established that India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
Also Read: Reliance AI Infrastructure in Jamnagar to Be Built in 24 Months: Akash Ambani
Ambani further stated that as a fundamental block, "We need to continue investing in AI infrastructure and AI data centers that are fully equipped to scale globally and support millions of users in India. At Jio, we are already doing this. We recently announced in Jamnagar that we are building our AI data center, which will have a gigawatt capacity."
He also stressed the importance of deep research and development. "We continue to invest in deep research and deep development that comes from it," he said, crediting Prime Minister Narendra Modi's vision for AI advancement through the AI mission.
Additionally, Ambani highlighted the need to invest in the right talent. "At Jio, we've already invested in our overall full-stack AI team, which is led by data scientists, researchers and engineers to be a thousand plus. The critical element in this, I feel, is also to foster new ideas and push the boundaries of development," he said, adding, "It's not too far away where we will have a groundbreaking idea that will have half a billion people on one platform coming out of India."
Perception of India as a Tech Laggard
Challenging the perception that India lags in technology adoption, Ambani pointed to the country's rapid digital transformation. "I think we've already showcased to the world in connectivity that we can be the leaders of technology, not just be fast followers."
He recalled India's shift in internet usage and speeds since 2015, saying, "...Back to 2015, which is just 10 years ago, Internet speeds were less than 1 MB in this country, whether it be on the mobile or at home. But today, everyone continues to enjoy very high data speeds. And so I think from an AI point of view is these three critical items that we need to focus on," Ambani said.
Replying to Ambani, Jain said, "But then when we grew up and we saw speeds in America, we'd be like, what the hell? Why is our speed, you know, why is our Internet speed like this in India? And why is this like, apparently 3G giving me less than 1 Mbps? And today I can say that when we travel, we laugh at the world. We literally were there and we are like, wait, you pay 60-70 USD a month for crappy Internet? Boss, come to India, just try Jio for a while and then you'll see what we have. So thank you for that. It's been an amazing thing to see all of India benefit from this."
Also Read: Akash Ambani Urges Rapid Adoption of AI and Data Centre Policy Reform in India
AI Taking Over Jobs
Responding to Jain's question about whether AI will eliminate many jobs, Ambani said, "I'm a firm believer that AI will transform jobs. Today we've seen AI take over our mundane tasks, our repetitive tasks."
"We at Jio are already embracing it," Ambani said, adding, "We've seen how the Internet can create new industries and jobs, but this time, we can prepare for it. And by preparing, I mean equipping ourselves with the right tools and insights to excel our impact, excel our businesses. For example, at Jio, today we monitor our network, which now covers about 95 percent of the living population of India."
Leveraging AI for Network Monitoring
Before that, we had to monitor networks across different systems. "Today we can proactively monitor and before even a customer has a bad experience, experience, we can predict it. And this is just not, you know, something that is, this is again, instead of deeping/delving deep into the data, figuring out what the consumer does, the patterns are being picked up by ML and AI engineers to kind of spend the time to actually solve the problem than finding out what the problem is. So that's one way that we prepared," Ambani explained.
AI in Education
Responding to Jain's idea of leveraging AI to solve education challenges in India, Ambani said, "For us, there are five layers that create a deep technology company. In this particular use case, the first thing to solve for in education is connectivity, which we pretty much solved in India."
Five Layers
Today, 1.5 million schools have connectivity. The second layer is computing—both data center computing and edge computing. At Jio, we are advancing in both across the nation. "After that comes the devices layer where we have to enable devices that can consume and actually deliver to the endpoint which in this case will be the student across, affordability, standpoints across," Ambani explained, referring to low-cost consumer devices.
"So it may be a cloud PC laptop or a low-cost laptop. It could be a smartphone tomorrow. It can be AR glasses that we can use to deliver these devices across to enhance the experience that
we can actually deliver. On top of that comes the content layer where content not in the form of videos but just educational content or anything. And then comes the intelligence layer. So these five layers put together, you know, enable us to become solving big problems like education," Ambani said.
"And finally the five layers that I spoke about, they need to create an absolutely seamless customer experience. If they do not create a seamless customer experience that is easy to adopt, easy to have, then it becomes very, very hard for us to solve large issues like the literacy rate and education in India. We're going to be doing that," he added.
He further said the same approach applies to agriculture, IoT devices, and healthcare. At Jio, we will work to scale solutions and impact millions of students and schools in India, Ambani said.
Aspiring Indian: AI
Jain expressed that the future of AI and technology in India will only be realized through full collaboration between public and private sectors.
Ambani responded, "We're extremely lucky to be under the visionary leadership of our Prime minister Narendra Modi on this." I think, it is the biggest blessing of our country to have a leader like him leading this mission. "Recently, at Parliament, he mentioned that AI does not only just start for artificial intelligence. It actually stands for Aspiring Indian."
Jio as a Platform Company
When asked whether Jio aims to become a platform company like Google, Amazon, Apple, or Meta, Ambani said, "At Jio, we also imagine ourselves to be platform companies. One of our biggest growth missions
of Jio and the way where we want to make impact is connecting each home. And in the home, we not only offer broadband connectivity but we offer our own OS which is a Tele OS that powers the large screen at the home."
"It is our vision to reach 100 million homes in the shortest amount of time possible. And I think we have now the technology and the demand to do so to enable that."
Replying to Ambani, Jain said, Take UPI, for example. We've almost started to take it for granted. But when you travel, you realize how far behind the world is in peer-to-peer payments. India has leapfrogged ahead.
"The way Jio took us from, you know, 3G, which wasn't working properly, to like 5G, which is the world's best…in a few years we leapfrogged that entire technology wave. And I can tell you all it's a small secret between 5,000 of us. But you know, this is not just a vision, this is actually something that, that Akash has running in his home. What you're talking, what he's talking about here is like, you know, have any of you all seen Iron man movie with Jarvis. Right? That's the kind of stuff they already have and they're working on to bring to our homes," Jain said, addressing the audience.
Also Read: Jio Platforms to Soon Launch Cloud-Based AI Personal Computer: Akash Ambani
Cloud-based AI PC
Ambani then spoke about enabling entrepreneurship. One possibility is GPU as a service—creating a foundation for developers to build on. Similarly, we will soon launch a cloud-based AI PC. You can read more about the Cloud-Based AI PC in the story linked above.
Jio Campus Open for Visits
He also announced that Jio's Mumbai campus, spanning 500 acres and home to 25,000 people, including 10,000–12,000 Jio employees, is now open for public visits. | 5 |
Artificial Intelligence (AI) is the biggest technological shift we have seen in our lifetime to date and will be the driving force behind India's economic growth, said Akash Ambani, Chairman of Reliance Jio Infocomm, at the Mumbai Tech Week held at Jio World Centre.
Also Read: Mumbai Set to Host Asia’s Largest AI Conference – MTW 2025
Investing in AI, Research, and Talent
"In my view, AI is the engine that will empower India to grow at 10 percent or double-digit growth numbers for the foreseeable future," Ambani stated, noting that "India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
To establish India as a global AI leader, Ambani outlined three key areas of focus: infrastructure, deep research and development, and talent investment.
Akash Ambani, Chairman of Reliance Jio Infocomm, spoke during a fireside chat with Dream11 CEO Harsh Jain at the Mumbai Tech Week held at Jio World Centre.
Jio’s Impact on India's Digital Transformation
Opening the conversation about Jio and its achievements in India, Dream11 CEO Harsh Jain said that Jio is literally at the forefront of technology in India. "Every single person here is thankful for finally getting us 5G Internet at the best speeds possible while we travel. I get like 150 Mbps in my car, on my phone, which is just phenomenal," Jain said.
"..We're all very fortunate to be in a country where our domestic product itself can reach 800 million Indians thanks to Jio. And it's a great time to be running a tech company and to have tech startups for all the entrepreneurs out there," Jain added.
India is a Leader in Adopting Technology
Responding to Jain's question about India still being seen as a laggard in technology, Akash Ambani replied, "I feel the contrary."
"Today, eight years after launching Jio, we've become the biggest data-consuming nation in the world. We've led that every, every single hour, average consumption per user is now significantly higher than anywhere else in the world, including China. So I think gone are the days that we should think about India from a tech laggard perspective. I think we have established that India is one of the forefront nations that can adopt technology and use technology for the benefit of the country."
Also Read: Reliance AI Infrastructure in Jamnagar to Be Built in 24 Months: Akash Ambani
Ambani further stated that as a fundamental block, "We need to continue investing in AI infrastructure and AI data centers that are fully equipped to scale globally and support millions of users in India. At Jio, we are already doing this. We recently announced in Jamnagar that we are building our AI data center, which will have a gigawatt capacity."
He also stressed the importance of deep research and development. "We continue to invest in deep research and deep development that comes from it," he said, crediting Prime Minister Narendra Modi's vision for AI advancement through the AI mission.
Additionally, Ambani highlighted the need to invest in the right talent. "At Jio, we've already invested in our overall full-stack AI team, which is led by data scientists, researchers and engineers to be a thousand plus. The critical element in this, I feel, is also to foster new ideas and push the boundaries of development," he said, adding, "It's not too far away where we will have a groundbreaking idea that will have half a billion people on one platform coming out of India."
Perception of India as a Tech Laggard
Challenging the perception that India lags in technology adoption, Ambani pointed to the country's rapid digital transformation. "I think we've already showcased to the world in connectivity that we can be the leaders of technology, not just be fast followers."
He recalled India's shift in internet usage and speeds since 2015, saying, "...Back to 2015, which is just 10 years ago, Internet speeds were less than 1 MB in this country, whether it be on the mobile or at home. But today, everyone continues to enjoy very high data speeds. And so I think from an AI point of view is these three critical items that we need to focus on," Ambani said.
Replying to Ambani, Jain said, "But then when we grew up and we saw speeds in America, we'd be like, what the hell? Why is our speed, you know, why is our Internet speed like this in India? And why is this like, apparently 3G giving me less than 1 Mbps? And today I can say that when we travel, we laugh at the world. We literally were there and we are like, wait, you pay 60-70 USD a month for crappy Internet? Boss, come to India, just try Jio for a while and then you'll see what we have. So thank you for that. It's been an amazing thing to see all of India benefit from this."
Also Read: Akash Ambani Urges Rapid Adoption of AI and Data Centre Policy Reform in India
AI Taking Over Jobs
Responding to Jain's question about whether AI will eliminate many jobs, Ambani said, "I'm a firm believer that AI will transform jobs. Today we've seen AI take over our mundane tasks, our repetitive tasks."
"We at Jio are already embracing it," Ambani said, adding, "We've seen how the Internet can create new industries and jobs, but this time, we can prepare for it. And by preparing, I mean equipping ourselves with the right tools and insights to excel our impact, excel our businesses. For example, at Jio, today we monitor our network, which now covers about 95 percent of the living population of India."
Leveraging AI for Network Monitoring
Before that, we had to monitor networks across different systems. "Today we can proactively monitor and before even a customer has a bad experience, experience, we can predict it. And this is just not, you know, something that is, this is again, instead of deeping/delving deep into the data, figuring out what the consumer does, the patterns are being picked up by ML and AI engineers to kind of spend the time to actually solve the problem than finding out what the problem is. So that's one way that we prepared," Ambani explained.
AI in Education
Responding to Jain's idea of leveraging AI to solve education challenges in India, Ambani said, "For us, there are five layers that create a deep technology company. In this particular use case, the first thing to solve for in education is connectivity, which we pretty much solved in India."
Five Layers
Today, 1.5 million schools have connectivity. The second layer is computing—both data center computing and edge computing. At Jio, we are advancing in both across the nation. "After that comes the devices layer where we have to enable devices that can consume and actually deliver to the endpoint which in this case will be the student across, affordability, standpoints across," Ambani explained, referring to low-cost consumer devices.
"So it may be a cloud PC laptop or a low-cost laptop. It could be a smartphone tomorrow. It can be AR glasses that we can use to deliver these devices across to enhance the experience that
we can actually deliver. On top of that comes the content layer where content not in the form of videos but just educational content or anything. And then comes the intelligence layer. So these five layers put together, you know, enable us to become solving big problems like education," Ambani said.
"And finally the five layers that I spoke about, they need to create an absolutely seamless customer experience. If they do not create a seamless customer experience that is easy to adopt, easy to have, then it becomes very, very hard for us to solve large issues like the literacy rate and education in India. We're going to be doing that," he added.
He further said the same approach applies to agriculture, IoT devices, and healthcare. At Jio, we will work to scale solutions and impact millions of students and schools in India, Ambani said.
Aspiring Indian: AI
Jain expressed that the future of AI and technology in India will only be realized through full collaboration between public and private sectors.
Ambani responded, "We're extremely lucky to be under the visionary leadership of our Prime minister Narendra Modi on this." I think, it is the biggest blessing of our country to have a leader like him leading this mission. "Recently, at Parliament, he mentioned that AI does not only just start for artificial intelligence. It actually stands for Aspiring Indian."
Jio as a Platform Company
When asked whether Jio aims to become a platform company like Google, Amazon, Apple, or Meta, Ambani said, "At Jio, we also imagine ourselves to be platform companies. One of our biggest growth missions
of Jio and the way where we want to make impact is connecting each home. And in the home, we not only offer broadband connectivity but we offer our own OS which is a Tele OS that powers the large screen at the home."
"It is our vision to reach 100 million homes in the shortest amount of time possible. And I think we have now the technology and the demand to do so to enable that."
Replying to Ambani, Jain said, Take UPI, for example. We've almost started to take it for granted. But when you travel, you realize how far behind the world is in peer-to-peer payments. India has leapfrogged ahead.
"The way Jio took us from, you know, 3G, which wasn't working properly, to like 5G, which is the world's best…in a few years we leapfrogged that entire technology wave. And I can tell you all it's a small secret between 5,000 of us. But you know, this is not just a vision, this is actually something that, that Akash has running in his home. What you're talking, what he's talking about here is like, you know, have any of you all seen Iron man movie with Jarvis. Right? That's the kind of stuff they already have and they're working on to bring to our homes," Jain said, addressing the audience.
Also Read: Jio Platforms to Soon Launch Cloud-Based AI Personal Computer: Akash Ambani
Cloud-based AI PC
Ambani then spoke about enabling entrepreneurship. One possibility is GPU as a service—creating a foundation for developers to build on. Similarly, we will soon launch a cloud-based AI PC. You can read more about the Cloud-Based AI PC in the story linked above.
Jio Campus Open for Visits
He also announced that Jio's Mumbai campus, spanning 500 acres and home to 25,000 people, including 10,000–12,000 Jio employees, is now open for public visits. | 5 |
Hamad Bin Khalifa Universitys (HBKU) Qatar Computing Research Institute (QCRI) and Google DeepMind co-hosted the 2025 Middle East and North Africa Machine Learning (MenaML) Winter School, gathering 141 local and international university students as well as industry professionals for a six-day program designed to equip them with critical skills in artificial intelligence (AI).The Winter School featured in-depth lectures, hands-on practical sessions and panel discussions that explored critical topics - including large language models, generative AI, deep learning, and more - as well as challenges and career opportunities for young professionals within the Arab world.It also explored the application of AI in diverse fields, including education, genomics, biology, and health, matching Qatars national challenge for AI practitioners to revolutionize these fields.Distinguished experts from QCRI, Google DeepMind, Carnegie Mellon University in Qatar, InstaDeep, and other regional organizations led these sessions, lending invaluable insights to the attendees and contributing to Qatars growing knowledge hub in AI and computer science.QCRI Executive Director Dr. Ahmed K. Elmagarmid commented: "Like our partners at Google DeepMind, QCRI focuses on stakeholders current and future needs in the field of computing. This natural partnership enabled us to help develop talent in AI based in Qatar and throughout the region, which in turn addresses national research priorities and contributes to the countrys knowledge-based economy."Participants also received opportunities for professional development, including mentorship sessions, networking gatherings, and a poster session where they could showcase their research to their peers and industry representatives.MenaML was sponsored and supported by a number of world-renowned companies, including Google DeepMind, Google Cloud, Google.org, Malomatia, Dell, PhazeRO, Scale, InstaDeep, Apple, and Qeen.AI, with many hosting booths and engaging directly with the students. MoroccoAI and inzva also served as community partners.The MenaML Winter School is designed to nurture the next generation of AI leaders and create a network of AI practitioners in the MENA region. The event fosters the exchange of knowledge, formation of new research collaborations, and development of local AI talent. ( | 5 |
Hamad Bin Khalifa Universitys (HBKU) Qatar Computing Research Institute (QCRI) and Google DeepMind co-hosted the 2025 Middle East and North Africa Machine Learning (MenaML) Winter School, gathering 141 local and international university students as well as industry professionals for a six-day program designed to equip them with critical skills in artificial intelligence (AI).The Winter School featured in-depth lectures, hands-on practical sessions and panel discussions that explored critical topics - including large language models, generative AI, deep learning, and more - as well as challenges and career opportunities for young professionals within the Arab world.It also explored the application of AI in diverse fields, including education, genomics, biology, and health, matching Qatars national challenge for AI practitioners to revolutionize these fields.Distinguished experts from QCRI, Google DeepMind, Carnegie Mellon University in Qatar, InstaDeep, and other regional organizations led these sessions, lending invaluable insights to the attendees and contributing to Qatars growing knowledge hub in AI and computer science.QCRI Executive Director Dr. Ahmed K. Elmagarmid commented: "Like our partners at Google DeepMind, QCRI focuses on stakeholders current and future needs in the field of computing. This natural partnership enabled us to help develop talent in AI based in Qatar and throughout the region, which in turn addresses national research priorities and contributes to the countrys knowledge-based economy."Participants also received opportunities for professional development, including mentorship sessions, networking gatherings, and a poster session where they could showcase their research to their peers and industry representatives.MenaML was sponsored and supported by a number of world-renowned companies, including Google DeepMind, Google Cloud, Google.org, Malomatia, Dell, PhazeRO, Scale, InstaDeep, Apple, and Qeen.AI, with many hosting booths and engaging directly with the students. MoroccoAI and inzva also served as community partners.The MenaML Winter School is designed to nurture the next generation of AI leaders and create a network of AI practitioners in the MENA region. The event fosters the exchange of knowledge, formation of new research collaborations, and development of local AI talent. ( | 5 |
Hamad Bin Khalifa Universitys (HBKU) Qatar Computing Research Institute (QCRI) and Google DeepMind co-hosted the 2025 Middle East and North Africa Machine Learning (MenaML) Winter School, gathering 141 local and international university students as well as industry professionals for a six-day program designed to equip them with critical skills in artificial intelligence (AI).The Winter School featured in-depth lectures, hands-on practical sessions and panel discussions that explored critical topics - including large language models, generative AI, deep learning, and more - as well as challenges and career opportunities for young professionals within the Arab world.It also explored the application of AI in diverse fields, including education, genomics, biology, and health, matching Qatars national challenge for AI practitioners to revolutionize these fields.Distinguished experts from QCRI, Google DeepMind, Carnegie Mellon University in Qatar, InstaDeep, and other regional organizations led these sessions, lending invaluable insights to the attendees and contributing to Qatars growing knowledge hub in AI and computer science.QCRI Executive Director Dr. Ahmed K. Elmagarmid commented: "Like our partners at Google DeepMind, QCRI focuses on stakeholders current and future needs in the field of computing. This natural partnership enabled us to help develop talent in AI based in Qatar and throughout the region, which in turn addresses national research priorities and contributes to the countrys knowledge-based economy."Participants also received opportunities for professional development, including mentorship sessions, networking gatherings, and a poster session where they could showcase their research to their peers and industry representatives.MenaML was sponsored and supported by a number of world-renowned companies, including Google DeepMind, Google Cloud, Google.org, Malomatia, Dell, PhazeRO, Scale, InstaDeep, Apple, and Qeen.AI, with many hosting booths and engaging directly with the students. MoroccoAI and inzva also served as community partners.The MenaML Winter School is designed to nurture the next generation of AI leaders and create a network of AI practitioners in the MENA region. The event fosters the exchange of knowledge, formation of new research collaborations, and development of local AI talent. ( | 5 |
Hamad Bin Khalifa Universitys (HBKU) Qatar Computing Research Institute (QCRI) and Google DeepMind co-hosted the 2025 Middle East and North Africa Machine Learning (MenaML) Winter School, gathering 141 local and international university students as well as industry professionals for a six-day program designed to equip them with critical skills in artificial intelligence (AI).The Winter School featured in-depth lectures, hands-on practical sessions and panel discussions that explored critical topics - including large language models, generative AI, deep learning, and more - as well as challenges and career opportunities for young professionals within the Arab world.It also explored the application of AI in diverse fields, including education, genomics, biology, and health, matching Qatars national challenge for AI practitioners to revolutionize these fields.Distinguished experts from QCRI, Google DeepMind, Carnegie Mellon University in Qatar, InstaDeep, and other regional organizations led these sessions, lending invaluable insights to the attendees and contributing to Qatars growing knowledge hub in AI and computer science.QCRI Executive Director Dr. Ahmed K. Elmagarmid commented: "Like our partners at Google DeepMind, QCRI focuses on stakeholders current and future needs in the field of computing. This natural partnership enabled us to help develop talent in AI based in Qatar and throughout the region, which in turn addresses national research priorities and contributes to the countrys knowledge-based economy."Participants also received opportunities for professional development, including mentorship sessions, networking gatherings, and a poster session where they could showcase their research to their peers and industry representatives.MenaML was sponsored and supported by a number of world-renowned companies, including Google DeepMind, Google Cloud, Google.org, Malomatia, Dell, PhazeRO, Scale, InstaDeep, Apple, and Qeen.AI, with many hosting booths and engaging directly with the students. MoroccoAI and inzva also served as community partners.The MenaML Winter School is designed to nurture the next generation of AI leaders and create a network of AI practitioners in the MENA region. The event fosters the exchange of knowledge, formation of new research collaborations, and development of local AI talent. ( | 5 |
Hamad Bin Khalifa Universitys (HBKU) Qatar Computing Research Institute (QCRI) and Google DeepMind co-hosted the 2025 Middle East and North Africa Machine Learning (MenaML) Winter School, gathering 141 local and international university students as well as industry professionals for a six-day program designed to equip them with critical skills in artificial intelligence (AI).The Winter School featured in-depth lectures, hands-on practical sessions and panel discussions that explored critical topics - including large language models, generative AI, deep learning, and more - as well as challenges and career opportunities for young professionals within the Arab world.It also explored the application of AI in diverse fields, including education, genomics, biology, and health, matching Qatars national challenge for AI practitioners to revolutionize these fields.Distinguished experts from QCRI, Google DeepMind, Carnegie Mellon University in Qatar, InstaDeep, and other regional organizations led these sessions, lending invaluable insights to the attendees and contributing to Qatars growing knowledge hub in AI and computer science.QCRI Executive Director Dr. Ahmed K. Elmagarmid commented: "Like our partners at Google DeepMind, QCRI focuses on stakeholders current and future needs in the field of computing. This natural partnership enabled us to help develop talent in AI based in Qatar and throughout the region, which in turn addresses national research priorities and contributes to the countrys knowledge-based economy."Participants also received opportunities for professional development, including mentorship sessions, networking gatherings, and a poster session where they could showcase their research to their peers and industry representatives.MenaML was sponsored and supported by a number of world-renowned companies, including Google DeepMind, Google Cloud, Google.org, Malomatia, Dell, PhazeRO, Scale, InstaDeep, Apple, and Qeen.AI, with many hosting booths and engaging directly with the students. MoroccoAI and inzva also served as community partners.The MenaML Winter School is designed to nurture the next generation of AI leaders and create a network of AI practitioners in the MENA region. The event fosters the exchange of knowledge, formation of new research collaborations, and development of local AI talent. ( | 5 |
This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 10, 2025 - March 16, 2025
US-based cybersecurity firm Palo Alto Networks’ Unit 42 researchers warn that the technology used by DeepSeek, a China-based artificial intelligence (AI) research organisation, is alarmingly vulnerable to jailbreaking compared with those of its peers and can produce nefarious content with little to no specialised knowledge or expertise.
Jailbreaking is used to bypass restrictions implemented in large language models (LLMs) to prevent them from generating malicious or prohibited content. These restrictions are commonly referred to as guardrails. If a straightforward request is put in as an LLM prompt, the guardrails will prevent the LLM from providing harmful content.
Unit 42 researchers recently uncovered two novel and effective jailbreaking techniques, Deceptive Delight and Bad Likert Judge. Given their success against other LLMs, the researchers tested these two techniques and a multi-stage jailbreaking technique called Crescendo against DeepSeek models.
While information on creating Molotov cocktails, data exfiltration tools and keyloggers are readily available online, LLMs with insufficient safety restrictions could lower the barrier to entry for malicious actors by compiling and presenting easily usable and actionable output. This assistance could greatly accelerate their operations.
Unit 42’s research findings show that these jailbreaking methods can elicit explicit guidance for malicious activities. These activities include data exfiltration tooling, keylogger creation and even instructions for incendiary devices, demonstrating the tangible security risks posed by this emerging class of attack.
In an email interview with Digital Edge, Philippa Cogswell, vice-president and managing partner at Unit 42 for Asia-Pacific and Japan at Palo Alto Networks, says it has conducted extensive testing on various LLMs, including those by other providers. However, its research found that DeepSeek is more susceptible to jailbreaking than other models.
“While we’ve successfully jailbroken numerous models, DeepSeek proved significantly easier to bypass. We achieved jailbreaks at a much faster rate given the absence of minimum guardrails designed to prevent the generation of malicious content. Our researchers were able to bypass its weak safeguards to generate harmful content, requiring little to no specialised knowledge or expertise,” she explains.
Cogswell adds that DeepSeek lacks the guardrails found in more established models like those of OpenAI, posing a greater concern due to its limited maturity and likely rushed-to-market release.
“While other models have undergone extensive red team exercises resulting in published research that provides clear insights into their security methodologies and frameworks, DeepSeek has not taken the time to do the due diligence to ensure proper guardrails were in place before going to market,” she notes.
AI is becoming an integral part of the workplace, with 75% of global knowledge workers and 84% of Malaysians already using the technology in their jobs. So, there is growing concern about how these systems handle sensitive information. Employees may unknowingly feed confidential data into AI models, raising serious security risks.
Jailbreaking tests are essential as they help uncover vulnerabilities in LLMs before bad actors can exploit these. The tests simulate real-world adversarial attacks to see if the AI can be tricked into generating harmful, biased or sensitive content that it was programmed to block.
“Identifying weaknesses early allows developers to strengthen safeguards, ensuring AI remains safe and trustworthy for users. LLMs are designed with safety measures that prevent them from generating responses that could be harmful or inappropriate. However, researchers have found ways to bypass these safeguards by manipulating how the AI processes language,” says Cogswell.
She explains that each technique exploits the model in a unique way. Deceptive Delight relies on cleverly worded prompts that mislead the AI into producing responses it would normally reject. Bad Likert Judge, on the other hand, manipulates the internal scoring system that the AI uses to assess whether a response is acceptable, essentially tricking it into approving restricted content. Crescendo takes a different approach by gradually escalating a prompt, sidestepping the AI’s filters that would block a more direct request.
“By understanding these methods, researchers and security teams can better prepare for potential attacks and improve AI defences accordingly,” she adds.
Taking proactive steps with LLMs
The key takeaway is that AI security should not be left solely in the hands of model providers. Organisations using LLMs need to take an active role in securing their AI applications, just as they would securing any other applications in their environment.
One of the most important steps is implementing additional guardrails at an organisational level, rather than relying entirely on the AI’s built-in safety measures, says Cogswell. This can include deploying internal filtering and monitoring tools that detect and prevent jailbreaking attempts.
“Beyond technical safeguards, human oversight remains critical. AI-generated content should be reviewed, especially in industries like finance, healthcare and cybersecurity, where accuracy and ethical considerations are paramount,” she says.
“Controlling access to LLMs is also essential — restricting who can use them, logging interactions and analysing usage patterns can help prevent misuse. Maintaining a zero-trust mindset is equally important: never trust and always verify, ensuring that every interaction and output is scrutinised for potential risks.”
Regular security testing should become standard practice. Just as cybersecurity teams perform penetration testing to find vulnerabilities in software, organisations should proactively assess their AI systems to stay ahead of evolving threats.
“Organisations are increasingly interested in a more structured, end-to-end assessment of their AI security, one that not only protects employees’ use of AI but also ensures responsible use of AI innovation and governance. Our insights are shaped from extensive threat research, results from real-world incident response cases and creative approaches from our consultants, such as red teamers, when testing AI applications,” says Cogswell.
AI is incredibly powerful, but we are still learning how to use it responsibly. Security is not just about technical barriers, but also about awareness and responsible use, she says. One of the most important steps organisations can take is educating employees on the risks associated with AI.
“Many security breaches happen because users unknowingly expose sensitive information or fall for social engineering tactics. Training staff to recognise threats like prompt injection, data leakage and misinformation can significantly reduce risks,” says Cogswell.
Another key principle is data protection. Users should be cautious about feeding confidential or proprietary information into public AI models as the way that data is stored and processed is not always transparent. For highly sensitive tasks, it is safer to use in-house or fine-tuned models that are specifically designed to protect organisational data.
Monitoring AI-generated content is also essential, where AI outputs should be regularly reviewed for bias, hallucinations or security risks. Organisations can implement automated scanning tools to flag potentially harmful responses before they cause damage, she says.
Finally, AI security should be a shared responsibility. Companies should collaborate with industry experts, participate in AI security research and stay informed about emerging threats. The more knowledge that is shared, the stronger the collective defence against AI-related risks.
“LLMs are here to stay, but ensuring their safe and responsible use requires a proactive, security-first mindset. By combining people, processes, technology and governance, organisations can harness AI’s full potential without compromising on security,” says Cogswell.
“As technology is ever-evolving, regular system updates and staying vigilant about the latest threats are essential to maintaining a robust defence against emerging risks.”
Save by subscribing to us for your print and/or digital copy. | 5 |
This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 10, 2025 - March 16, 2025
US-based cybersecurity firm Palo Alto Networks’ Unit 42 researchers warn that the technology used by DeepSeek, a China-based artificial intelligence (AI) research organisation, is alarmingly vulnerable to jailbreaking compared with those of its peers and can produce nefarious content with little to no specialised knowledge or expertise.
Jailbreaking is used to bypass restrictions implemented in large language models (LLMs) to prevent them from generating malicious or prohibited content. These restrictions are commonly referred to as guardrails. If a straightforward request is put in as an LLM prompt, the guardrails will prevent the LLM from providing harmful content.
Unit 42 researchers recently uncovered two novel and effective jailbreaking techniques, Deceptive Delight and Bad Likert Judge. Given their success against other LLMs, the researchers tested these two techniques and a multi-stage jailbreaking technique called Crescendo against DeepSeek models.
While information on creating Molotov cocktails, data exfiltration tools and keyloggers are readily available online, LLMs with insufficient safety restrictions could lower the barrier to entry for malicious actors by compiling and presenting easily usable and actionable output. This assistance could greatly accelerate their operations.
Unit 42’s research findings show that these jailbreaking methods can elicit explicit guidance for malicious activities. These activities include data exfiltration tooling, keylogger creation and even instructions for incendiary devices, demonstrating the tangible security risks posed by this emerging class of attack.
In an email interview with Digital Edge, Philippa Cogswell, vice-president and managing partner at Unit 42 for Asia-Pacific and Japan at Palo Alto Networks, says it has conducted extensive testing on various LLMs, including those by other providers. However, its research found that DeepSeek is more susceptible to jailbreaking than other models.
“While we’ve successfully jailbroken numerous models, DeepSeek proved significantly easier to bypass. We achieved jailbreaks at a much faster rate given the absence of minimum guardrails designed to prevent the generation of malicious content. Our researchers were able to bypass its weak safeguards to generate harmful content, requiring little to no specialised knowledge or expertise,” she explains.
Cogswell adds that DeepSeek lacks the guardrails found in more established models like those of OpenAI, posing a greater concern due to its limited maturity and likely rushed-to-market release.
“While other models have undergone extensive red team exercises resulting in published research that provides clear insights into their security methodologies and frameworks, DeepSeek has not taken the time to do the due diligence to ensure proper guardrails were in place before going to market,” she notes.
AI is becoming an integral part of the workplace, with 75% of global knowledge workers and 84% of Malaysians already using the technology in their jobs. So, there is growing concern about how these systems handle sensitive information. Employees may unknowingly feed confidential data into AI models, raising serious security risks.
Jailbreaking tests are essential as they help uncover vulnerabilities in LLMs before bad actors can exploit these. The tests simulate real-world adversarial attacks to see if the AI can be tricked into generating harmful, biased or sensitive content that it was programmed to block.
“Identifying weaknesses early allows developers to strengthen safeguards, ensuring AI remains safe and trustworthy for users. LLMs are designed with safety measures that prevent them from generating responses that could be harmful or inappropriate. However, researchers have found ways to bypass these safeguards by manipulating how the AI processes language,” says Cogswell.
She explains that each technique exploits the model in a unique way. Deceptive Delight relies on cleverly worded prompts that mislead the AI into producing responses it would normally reject. Bad Likert Judge, on the other hand, manipulates the internal scoring system that the AI uses to assess whether a response is acceptable, essentially tricking it into approving restricted content. Crescendo takes a different approach by gradually escalating a prompt, sidestepping the AI’s filters that would block a more direct request.
“By understanding these methods, researchers and security teams can better prepare for potential attacks and improve AI defences accordingly,” she adds.
Taking proactive steps with LLMs
The key takeaway is that AI security should not be left solely in the hands of model providers. Organisations using LLMs need to take an active role in securing their AI applications, just as they would securing any other applications in their environment.
One of the most important steps is implementing additional guardrails at an organisational level, rather than relying entirely on the AI’s built-in safety measures, says Cogswell. This can include deploying internal filtering and monitoring tools that detect and prevent jailbreaking attempts.
“Beyond technical safeguards, human oversight remains critical. AI-generated content should be reviewed, especially in industries like finance, healthcare and cybersecurity, where accuracy and ethical considerations are paramount,” she says.
“Controlling access to LLMs is also essential — restricting who can use them, logging interactions and analysing usage patterns can help prevent misuse. Maintaining a zero-trust mindset is equally important: never trust and always verify, ensuring that every interaction and output is scrutinised for potential risks.”
Regular security testing should become standard practice. Just as cybersecurity teams perform penetration testing to find vulnerabilities in software, organisations should proactively assess their AI systems to stay ahead of evolving threats.
“Organisations are increasingly interested in a more structured, end-to-end assessment of their AI security, one that not only protects employees’ use of AI but also ensures responsible use of AI innovation and governance. Our insights are shaped from extensive threat research, results from real-world incident response cases and creative approaches from our consultants, such as red teamers, when testing AI applications,” says Cogswell.
AI is incredibly powerful, but we are still learning how to use it responsibly. Security is not just about technical barriers, but also about awareness and responsible use, she says. One of the most important steps organisations can take is educating employees on the risks associated with AI.
“Many security breaches happen because users unknowingly expose sensitive information or fall for social engineering tactics. Training staff to recognise threats like prompt injection, data leakage and misinformation can significantly reduce risks,” says Cogswell.
Another key principle is data protection. Users should be cautious about feeding confidential or proprietary information into public AI models as the way that data is stored and processed is not always transparent. For highly sensitive tasks, it is safer to use in-house or fine-tuned models that are specifically designed to protect organisational data.
Monitoring AI-generated content is also essential, where AI outputs should be regularly reviewed for bias, hallucinations or security risks. Organisations can implement automated scanning tools to flag potentially harmful responses before they cause damage, she says.
Finally, AI security should be a shared responsibility. Companies should collaborate with industry experts, participate in AI security research and stay informed about emerging threats. The more knowledge that is shared, the stronger the collective defence against AI-related risks.
“LLMs are here to stay, but ensuring their safe and responsible use requires a proactive, security-first mindset. By combining people, processes, technology and governance, organisations can harness AI’s full potential without compromising on security,” says Cogswell.
“As technology is ever-evolving, regular system updates and staying vigilant about the latest threats are essential to maintaining a robust defence against emerging risks.”
Save by subscribing to us for your print and/or digital copy. | 5 |
This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 10, 2025 - March 16, 2025
US-based cybersecurity firm Palo Alto Networks’ Unit 42 researchers warn that the technology used by DeepSeek, a China-based artificial intelligence (AI) research organisation, is alarmingly vulnerable to jailbreaking compared with those of its peers and can produce nefarious content with little to no specialised knowledge or expertise.
Jailbreaking is used to bypass restrictions implemented in large language models (LLMs) to prevent them from generating malicious or prohibited content. These restrictions are commonly referred to as guardrails. If a straightforward request is put in as an LLM prompt, the guardrails will prevent the LLM from providing harmful content.
Unit 42 researchers recently uncovered two novel and effective jailbreaking techniques, Deceptive Delight and Bad Likert Judge. Given their success against other LLMs, the researchers tested these two techniques and a multi-stage jailbreaking technique called Crescendo against DeepSeek models.
While information on creating Molotov cocktails, data exfiltration tools and keyloggers are readily available online, LLMs with insufficient safety restrictions could lower the barrier to entry for malicious actors by compiling and presenting easily usable and actionable output. This assistance could greatly accelerate their operations.
Unit 42’s research findings show that these jailbreaking methods can elicit explicit guidance for malicious activities. These activities include data exfiltration tooling, keylogger creation and even instructions for incendiary devices, demonstrating the tangible security risks posed by this emerging class of attack.
In an email interview with Digital Edge, Philippa Cogswell, vice-president and managing partner at Unit 42 for Asia-Pacific and Japan at Palo Alto Networks, says it has conducted extensive testing on various LLMs, including those by other providers. However, its research found that DeepSeek is more susceptible to jailbreaking than other models.
“While we’ve successfully jailbroken numerous models, DeepSeek proved significantly easier to bypass. We achieved jailbreaks at a much faster rate given the absence of minimum guardrails designed to prevent the generation of malicious content. Our researchers were able to bypass its weak safeguards to generate harmful content, requiring little to no specialised knowledge or expertise,” she explains.
Cogswell adds that DeepSeek lacks the guardrails found in more established models like those of OpenAI, posing a greater concern due to its limited maturity and likely rushed-to-market release.
“While other models have undergone extensive red team exercises resulting in published research that provides clear insights into their security methodologies and frameworks, DeepSeek has not taken the time to do the due diligence to ensure proper guardrails were in place before going to market,” she notes.
AI is becoming an integral part of the workplace, with 75% of global knowledge workers and 84% of Malaysians already using the technology in their jobs. So, there is growing concern about how these systems handle sensitive information. Employees may unknowingly feed confidential data into AI models, raising serious security risks.
Jailbreaking tests are essential as they help uncover vulnerabilities in LLMs before bad actors can exploit these. The tests simulate real-world adversarial attacks to see if the AI can be tricked into generating harmful, biased or sensitive content that it was programmed to block.
“Identifying weaknesses early allows developers to strengthen safeguards, ensuring AI remains safe and trustworthy for users. LLMs are designed with safety measures that prevent them from generating responses that could be harmful or inappropriate. However, researchers have found ways to bypass these safeguards by manipulating how the AI processes language,” says Cogswell.
She explains that each technique exploits the model in a unique way. Deceptive Delight relies on cleverly worded prompts that mislead the AI into producing responses it would normally reject. Bad Likert Judge, on the other hand, manipulates the internal scoring system that the AI uses to assess whether a response is acceptable, essentially tricking it into approving restricted content. Crescendo takes a different approach by gradually escalating a prompt, sidestepping the AI’s filters that would block a more direct request.
“By understanding these methods, researchers and security teams can better prepare for potential attacks and improve AI defences accordingly,” she adds.
Taking proactive steps with LLMs
The key takeaway is that AI security should not be left solely in the hands of model providers. Organisations using LLMs need to take an active role in securing their AI applications, just as they would securing any other applications in their environment.
One of the most important steps is implementing additional guardrails at an organisational level, rather than relying entirely on the AI’s built-in safety measures, says Cogswell. This can include deploying internal filtering and monitoring tools that detect and prevent jailbreaking attempts.
“Beyond technical safeguards, human oversight remains critical. AI-generated content should be reviewed, especially in industries like finance, healthcare and cybersecurity, where accuracy and ethical considerations are paramount,” she says.
“Controlling access to LLMs is also essential — restricting who can use them, logging interactions and analysing usage patterns can help prevent misuse. Maintaining a zero-trust mindset is equally important: never trust and always verify, ensuring that every interaction and output is scrutinised for potential risks.”
Regular security testing should become standard practice. Just as cybersecurity teams perform penetration testing to find vulnerabilities in software, organisations should proactively assess their AI systems to stay ahead of evolving threats.
“Organisations are increasingly interested in a more structured, end-to-end assessment of their AI security, one that not only protects employees’ use of AI but also ensures responsible use of AI innovation and governance. Our insights are shaped from extensive threat research, results from real-world incident response cases and creative approaches from our consultants, such as red teamers, when testing AI applications,” says Cogswell.
AI is incredibly powerful, but we are still learning how to use it responsibly. Security is not just about technical barriers, but also about awareness and responsible use, she says. One of the most important steps organisations can take is educating employees on the risks associated with AI.
“Many security breaches happen because users unknowingly expose sensitive information or fall for social engineering tactics. Training staff to recognise threats like prompt injection, data leakage and misinformation can significantly reduce risks,” says Cogswell.
Another key principle is data protection. Users should be cautious about feeding confidential or proprietary information into public AI models as the way that data is stored and processed is not always transparent. For highly sensitive tasks, it is safer to use in-house or fine-tuned models that are specifically designed to protect organisational data.
Monitoring AI-generated content is also essential, where AI outputs should be regularly reviewed for bias, hallucinations or security risks. Organisations can implement automated scanning tools to flag potentially harmful responses before they cause damage, she says.
Finally, AI security should be a shared responsibility. Companies should collaborate with industry experts, participate in AI security research and stay informed about emerging threats. The more knowledge that is shared, the stronger the collective defence against AI-related risks.
“LLMs are here to stay, but ensuring their safe and responsible use requires a proactive, security-first mindset. By combining people, processes, technology and governance, organisations can harness AI’s full potential without compromising on security,” says Cogswell.
“As technology is ever-evolving, regular system updates and staying vigilant about the latest threats are essential to maintaining a robust defence against emerging risks.”
Save by subscribing to us for your print and/or digital copy. | 5 |
This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 10, 2025 - March 16, 2025
US-based cybersecurity firm Palo Alto Networks’ Unit 42 researchers warn that the technology used by DeepSeek, a China-based artificial intelligence (AI) research organisation, is alarmingly vulnerable to jailbreaking compared with those of its peers and can produce nefarious content with little to no specialised knowledge or expertise.
Jailbreaking is used to bypass restrictions implemented in large language models (LLMs) to prevent them from generating malicious or prohibited content. These restrictions are commonly referred to as guardrails. If a straightforward request is put in as an LLM prompt, the guardrails will prevent the LLM from providing harmful content.
Unit 42 researchers recently uncovered two novel and effective jailbreaking techniques, Deceptive Delight and Bad Likert Judge. Given their success against other LLMs, the researchers tested these two techniques and a multi-stage jailbreaking technique called Crescendo against DeepSeek models.
While information on creating Molotov cocktails, data exfiltration tools and keyloggers are readily available online, LLMs with insufficient safety restrictions could lower the barrier to entry for malicious actors by compiling and presenting easily usable and actionable output. This assistance could greatly accelerate their operations.
Unit 42’s research findings show that these jailbreaking methods can elicit explicit guidance for malicious activities. These activities include data exfiltration tooling, keylogger creation and even instructions for incendiary devices, demonstrating the tangible security risks posed by this emerging class of attack.
In an email interview with Digital Edge, Philippa Cogswell, vice-president and managing partner at Unit 42 for Asia-Pacific and Japan at Palo Alto Networks, says it has conducted extensive testing on various LLMs, including those by other providers. However, its research found that DeepSeek is more susceptible to jailbreaking than other models.
“While we’ve successfully jailbroken numerous models, DeepSeek proved significantly easier to bypass. We achieved jailbreaks at a much faster rate given the absence of minimum guardrails designed to prevent the generation of malicious content. Our researchers were able to bypass its weak safeguards to generate harmful content, requiring little to no specialised knowledge or expertise,” she explains.
Cogswell adds that DeepSeek lacks the guardrails found in more established models like those of OpenAI, posing a greater concern due to its limited maturity and likely rushed-to-market release.
“While other models have undergone extensive red team exercises resulting in published research that provides clear insights into their security methodologies and frameworks, DeepSeek has not taken the time to do the due diligence to ensure proper guardrails were in place before going to market,” she notes.
AI is becoming an integral part of the workplace, with 75% of global knowledge workers and 84% of Malaysians already using the technology in their jobs. So, there is growing concern about how these systems handle sensitive information. Employees may unknowingly feed confidential data into AI models, raising serious security risks.
Jailbreaking tests are essential as they help uncover vulnerabilities in LLMs before bad actors can exploit these. The tests simulate real-world adversarial attacks to see if the AI can be tricked into generating harmful, biased or sensitive content that it was programmed to block.
“Identifying weaknesses early allows developers to strengthen safeguards, ensuring AI remains safe and trustworthy for users. LLMs are designed with safety measures that prevent them from generating responses that could be harmful or inappropriate. However, researchers have found ways to bypass these safeguards by manipulating how the AI processes language,” says Cogswell.
She explains that each technique exploits the model in a unique way. Deceptive Delight relies on cleverly worded prompts that mislead the AI into producing responses it would normally reject. Bad Likert Judge, on the other hand, manipulates the internal scoring system that the AI uses to assess whether a response is acceptable, essentially tricking it into approving restricted content. Crescendo takes a different approach by gradually escalating a prompt, sidestepping the AI’s filters that would block a more direct request.
“By understanding these methods, researchers and security teams can better prepare for potential attacks and improve AI defences accordingly,” she adds.
Taking proactive steps with LLMs
The key takeaway is that AI security should not be left solely in the hands of model providers. Organisations using LLMs need to take an active role in securing their AI applications, just as they would securing any other applications in their environment.
One of the most important steps is implementing additional guardrails at an organisational level, rather than relying entirely on the AI’s built-in safety measures, says Cogswell. This can include deploying internal filtering and monitoring tools that detect and prevent jailbreaking attempts.
“Beyond technical safeguards, human oversight remains critical. AI-generated content should be reviewed, especially in industries like finance, healthcare and cybersecurity, where accuracy and ethical considerations are paramount,” she says.
“Controlling access to LLMs is also essential — restricting who can use them, logging interactions and analysing usage patterns can help prevent misuse. Maintaining a zero-trust mindset is equally important: never trust and always verify, ensuring that every interaction and output is scrutinised for potential risks.”
Regular security testing should become standard practice. Just as cybersecurity teams perform penetration testing to find vulnerabilities in software, organisations should proactively assess their AI systems to stay ahead of evolving threats.
“Organisations are increasingly interested in a more structured, end-to-end assessment of their AI security, one that not only protects employees’ use of AI but also ensures responsible use of AI innovation and governance. Our insights are shaped from extensive threat research, results from real-world incident response cases and creative approaches from our consultants, such as red teamers, when testing AI applications,” says Cogswell.
AI is incredibly powerful, but we are still learning how to use it responsibly. Security is not just about technical barriers, but also about awareness and responsible use, she says. One of the most important steps organisations can take is educating employees on the risks associated with AI.
“Many security breaches happen because users unknowingly expose sensitive information or fall for social engineering tactics. Training staff to recognise threats like prompt injection, data leakage and misinformation can significantly reduce risks,” says Cogswell.
Another key principle is data protection. Users should be cautious about feeding confidential or proprietary information into public AI models as the way that data is stored and processed is not always transparent. For highly sensitive tasks, it is safer to use in-house or fine-tuned models that are specifically designed to protect organisational data.
Monitoring AI-generated content is also essential, where AI outputs should be regularly reviewed for bias, hallucinations or security risks. Organisations can implement automated scanning tools to flag potentially harmful responses before they cause damage, she says.
Finally, AI security should be a shared responsibility. Companies should collaborate with industry experts, participate in AI security research and stay informed about emerging threats. The more knowledge that is shared, the stronger the collective defence against AI-related risks.
“LLMs are here to stay, but ensuring their safe and responsible use requires a proactive, security-first mindset. By combining people, processes, technology and governance, organisations can harness AI’s full potential without compromising on security,” says Cogswell.
“As technology is ever-evolving, regular system updates and staying vigilant about the latest threats are essential to maintaining a robust defence against emerging risks.”
Save by subscribing to us for your print and/or digital copy. | 5 |
This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 10, 2025 - March 16, 2025
US-based cybersecurity firm Palo Alto Networks’ Unit 42 researchers warn that the technology used by DeepSeek, a China-based artificial intelligence (AI) research organisation, is alarmingly vulnerable to jailbreaking compared with those of its peers and can produce nefarious content with little to no specialised knowledge or expertise.
Jailbreaking is used to bypass restrictions implemented in large language models (LLMs) to prevent them from generating malicious or prohibited content. These restrictions are commonly referred to as guardrails. If a straightforward request is put in as an LLM prompt, the guardrails will prevent the LLM from providing harmful content.
Unit 42 researchers recently uncovered two novel and effective jailbreaking techniques, Deceptive Delight and Bad Likert Judge. Given their success against other LLMs, the researchers tested these two techniques and a multi-stage jailbreaking technique called Crescendo against DeepSeek models.
While information on creating Molotov cocktails, data exfiltration tools and keyloggers are readily available online, LLMs with insufficient safety restrictions could lower the barrier to entry for malicious actors by compiling and presenting easily usable and actionable output. This assistance could greatly accelerate their operations.
Unit 42’s research findings show that these jailbreaking methods can elicit explicit guidance for malicious activities. These activities include data exfiltration tooling, keylogger creation and even instructions for incendiary devices, demonstrating the tangible security risks posed by this emerging class of attack.
In an email interview with Digital Edge, Philippa Cogswell, vice-president and managing partner at Unit 42 for Asia-Pacific and Japan at Palo Alto Networks, says it has conducted extensive testing on various LLMs, including those by other providers. However, its research found that DeepSeek is more susceptible to jailbreaking than other models.
“While we’ve successfully jailbroken numerous models, DeepSeek proved significantly easier to bypass. We achieved jailbreaks at a much faster rate given the absence of minimum guardrails designed to prevent the generation of malicious content. Our researchers were able to bypass its weak safeguards to generate harmful content, requiring little to no specialised knowledge or expertise,” she explains.
Cogswell adds that DeepSeek lacks the guardrails found in more established models like those of OpenAI, posing a greater concern due to its limited maturity and likely rushed-to-market release.
“While other models have undergone extensive red team exercises resulting in published research that provides clear insights into their security methodologies and frameworks, DeepSeek has not taken the time to do the due diligence to ensure proper guardrails were in place before going to market,” she notes.
AI is becoming an integral part of the workplace, with 75% of global knowledge workers and 84% of Malaysians already using the technology in their jobs. So, there is growing concern about how these systems handle sensitive information. Employees may unknowingly feed confidential data into AI models, raising serious security risks.
Jailbreaking tests are essential as they help uncover vulnerabilities in LLMs before bad actors can exploit these. The tests simulate real-world adversarial attacks to see if the AI can be tricked into generating harmful, biased or sensitive content that it was programmed to block.
“Identifying weaknesses early allows developers to strengthen safeguards, ensuring AI remains safe and trustworthy for users. LLMs are designed with safety measures that prevent them from generating responses that could be harmful or inappropriate. However, researchers have found ways to bypass these safeguards by manipulating how the AI processes language,” says Cogswell.
She explains that each technique exploits the model in a unique way. Deceptive Delight relies on cleverly worded prompts that mislead the AI into producing responses it would normally reject. Bad Likert Judge, on the other hand, manipulates the internal scoring system that the AI uses to assess whether a response is acceptable, essentially tricking it into approving restricted content. Crescendo takes a different approach by gradually escalating a prompt, sidestepping the AI’s filters that would block a more direct request.
“By understanding these methods, researchers and security teams can better prepare for potential attacks and improve AI defences accordingly,” she adds.
Taking proactive steps with LLMs
The key takeaway is that AI security should not be left solely in the hands of model providers. Organisations using LLMs need to take an active role in securing their AI applications, just as they would securing any other applications in their environment.
One of the most important steps is implementing additional guardrails at an organisational level, rather than relying entirely on the AI’s built-in safety measures, says Cogswell. This can include deploying internal filtering and monitoring tools that detect and prevent jailbreaking attempts.
“Beyond technical safeguards, human oversight remains critical. AI-generated content should be reviewed, especially in industries like finance, healthcare and cybersecurity, where accuracy and ethical considerations are paramount,” she says.
“Controlling access to LLMs is also essential — restricting who can use them, logging interactions and analysing usage patterns can help prevent misuse. Maintaining a zero-trust mindset is equally important: never trust and always verify, ensuring that every interaction and output is scrutinised for potential risks.”
Regular security testing should become standard practice. Just as cybersecurity teams perform penetration testing to find vulnerabilities in software, organisations should proactively assess their AI systems to stay ahead of evolving threats.
“Organisations are increasingly interested in a more structured, end-to-end assessment of their AI security, one that not only protects employees’ use of AI but also ensures responsible use of AI innovation and governance. Our insights are shaped from extensive threat research, results from real-world incident response cases and creative approaches from our consultants, such as red teamers, when testing AI applications,” says Cogswell.
AI is incredibly powerful, but we are still learning how to use it responsibly. Security is not just about technical barriers, but also about awareness and responsible use, she says. One of the most important steps organisations can take is educating employees on the risks associated with AI.
“Many security breaches happen because users unknowingly expose sensitive information or fall for social engineering tactics. Training staff to recognise threats like prompt injection, data leakage and misinformation can significantly reduce risks,” says Cogswell.
Another key principle is data protection. Users should be cautious about feeding confidential or proprietary information into public AI models as the way that data is stored and processed is not always transparent. For highly sensitive tasks, it is safer to use in-house or fine-tuned models that are specifically designed to protect organisational data.
Monitoring AI-generated content is also essential, where AI outputs should be regularly reviewed for bias, hallucinations or security risks. Organisations can implement automated scanning tools to flag potentially harmful responses before they cause damage, she says.
Finally, AI security should be a shared responsibility. Companies should collaborate with industry experts, participate in AI security research and stay informed about emerging threats. The more knowledge that is shared, the stronger the collective defence against AI-related risks.
“LLMs are here to stay, but ensuring their safe and responsible use requires a proactive, security-first mindset. By combining people, processes, technology and governance, organisations can harness AI’s full potential without compromising on security,” says Cogswell.
“As technology is ever-evolving, regular system updates and staying vigilant about the latest threats are essential to maintaining a robust defence against emerging risks.”
Save by subscribing to us for your print and/or digital copy. | 5 |
사진 확대
Artificial intelligence (AI) technology is becoming a key driver of industries and economies in the 21st century, and competition for supremacy by countries around the world is intensifying day by day. In particular, not long ago, Chinese AI start-up "DeepSeek" launched a new AI model that is almost equivalent to the performance of its competitors at a much lower cost, shocking the United States and the rest of the world. The 'deep shock', which is recognized as a groundbreaking achievement that is called the 'Sputnik moment of AI', is recognized as a great threat to the United States, which has led the innovation of the AI industry.Amid intensifying competition for global hegemony in the AI industry, mainly in the United States and China, South Korea launched the National AI Committee headed by the president in September last year and entered full-fledged competition with the vision of becoming one of the "three AI powers." However, even in the AI industry, the gap between the metropolitan area and provinces is gradually widening. If this imbalance is not resolved, Korea will be eliminated from global competition.Pohang, which is transforming from a steel industry city into a high-tech new industrial city such as a secondary battery, has a wide range of conditions for the AI industry to grow rapidly. In terms of balanced regional development, it also fits the optimal location of the AI industry. This is why Pohang is actively seeking to attract a national AI computing center that the government is promoting. The National AI Computing Center, which will accelerate the leap forward as one of the three AI powerhouses, is a base infrastructure that combines advanced AI semiconductors and dedicated programs to produce useful results needed by companies and research institutes from vast amounts of data, and the government plans to build it in non-metropolitan areas. Pohang has the optimal requirements for an AI computing center to be located, including abundant high-tech research facilities, talent, and stable power supply and demand.First of all, 4th generation radiation accelerators, POSTECH AI Research Institute, Apple R&D Support Center, and Robot Convergence Research Institute are integrated, making it easy to develop related technologies through research activities and having an environment where large-scale data processing and analysis are essential for the development of AI technology. In addition, there are plenty of master's and doctoral-level researchers in the AI field, centering on POSTECH and Handong University.As nuclear power plants are concentrated nearby, it is possible to supply stable power, which is important for fostering the AI industry, and if new and renewable energy generation facilities such as hydrogen fuel cells and offshore wind power are expanded, it is possible to supply low-cost power using the Special Act on Promotion of Distributed Energy.In particular, Pohang, which is leading innovation in new industries such as secondary batteries, bio, and hydrogen along with the steel industry, a traditional manufacturing industry, has optimal conditions to build a smart manufacturing ecosystem that combines AI technology. In addition, AI computing-related infrastructure such as the 1.52 trillion won global data center campus development project, the establishment of an AI open innovation center, and the establishment of an AI accelerator center is also progressing smoothly.Pohang's development of the AI industry will lead to innovation in the entire industry through convergence with manufacturing, Korea's main industry, beyond just regional development. Pohang City intends to begin its journey to become a Sovereign (sovereign) AI hub city in Korea based on specific strategies, action plans, strong industrial base and human resources. I sincerely hope that the starting point will be the establishment of Pohang, an AI computing center, with 500,000 citizens.[Lee Kang Deok, Mayor of Pohang] | 5 |
사진 확대
Artificial intelligence (AI) technology is becoming a key driver of industries and economies in the 21st century, and competition for supremacy by countries around the world is intensifying day by day. In particular, not long ago, Chinese AI start-up "DeepSeek" launched a new AI model that is almost equivalent to the performance of its competitors at a much lower cost, shocking the United States and the rest of the world. The 'deep shock', which is recognized as a groundbreaking achievement that is called the 'Sputnik moment of AI', is recognized as a great threat to the United States, which has led the innovation of the AI industry.Amid intensifying competition for global hegemony in the AI industry, mainly in the United States and China, South Korea launched the National AI Committee headed by the president in September last year and entered full-fledged competition with the vision of becoming one of the "three AI powers." However, even in the AI industry, the gap between the metropolitan area and provinces is gradually widening. If this imbalance is not resolved, Korea will be eliminated from global competition.Pohang, which is transforming from a steel industry city into a high-tech new industrial city such as a secondary battery, has a wide range of conditions for the AI industry to grow rapidly. In terms of balanced regional development, it also fits the optimal location of the AI industry. This is why Pohang is actively seeking to attract a national AI computing center that the government is promoting. The National AI Computing Center, which will accelerate the leap forward as one of the three AI powerhouses, is a base infrastructure that combines advanced AI semiconductors and dedicated programs to produce useful results needed by companies and research institutes from vast amounts of data, and the government plans to build it in non-metropolitan areas. Pohang has the optimal requirements for an AI computing center to be located, including abundant high-tech research facilities, talent, and stable power supply and demand.First of all, 4th generation radiation accelerators, POSTECH AI Research Institute, Apple R&D Support Center, and Robot Convergence Research Institute are integrated, making it easy to develop related technologies through research activities and having an environment where large-scale data processing and analysis are essential for the development of AI technology. In addition, there are plenty of master's and doctoral-level researchers in the AI field, centering on POSTECH and Handong University.As nuclear power plants are concentrated nearby, it is possible to supply stable power, which is important for fostering the AI industry, and if new and renewable energy generation facilities such as hydrogen fuel cells and offshore wind power are expanded, it is possible to supply low-cost power using the Special Act on Promotion of Distributed Energy.In particular, Pohang, which is leading innovation in new industries such as secondary batteries, bio, and hydrogen along with the steel industry, a traditional manufacturing industry, has optimal conditions to build a smart manufacturing ecosystem that combines AI technology. In addition, AI computing-related infrastructure such as the 1.52 trillion won global data center campus development project, the establishment of an AI open innovation center, and the establishment of an AI accelerator center is also progressing smoothly.Pohang's development of the AI industry will lead to innovation in the entire industry through convergence with manufacturing, Korea's main industry, beyond just regional development. Pohang City intends to begin its journey to become a Sovereign (sovereign) AI hub city in Korea based on specific strategies, action plans, strong industrial base and human resources. I sincerely hope that the starting point will be the establishment of Pohang, an AI computing center, with 500,000 citizens.[Lee Kang Deok, Mayor of Pohang] | 5 |
사진 확대
Artificial intelligence (AI) technology is becoming a key driver of industries and economies in the 21st century, and competition for supremacy by countries around the world is intensifying day by day. In particular, not long ago, Chinese AI start-up "DeepSeek" launched a new AI model that is almost equivalent to the performance of its competitors at a much lower cost, shocking the United States and the rest of the world. The 'deep shock', which is recognized as a groundbreaking achievement that is called the 'Sputnik moment of AI', is recognized as a great threat to the United States, which has led the innovation of the AI industry.Amid intensifying competition for global hegemony in the AI industry, mainly in the United States and China, South Korea launched the National AI Committee headed by the president in September last year and entered full-fledged competition with the vision of becoming one of the "three AI powers." However, even in the AI industry, the gap between the metropolitan area and provinces is gradually widening. If this imbalance is not resolved, Korea will be eliminated from global competition.Pohang, which is transforming from a steel industry city into a high-tech new industrial city such as a secondary battery, has a wide range of conditions for the AI industry to grow rapidly. In terms of balanced regional development, it also fits the optimal location of the AI industry. This is why Pohang is actively seeking to attract a national AI computing center that the government is promoting. The National AI Computing Center, which will accelerate the leap forward as one of the three AI powerhouses, is a base infrastructure that combines advanced AI semiconductors and dedicated programs to produce useful results needed by companies and research institutes from vast amounts of data, and the government plans to build it in non-metropolitan areas. Pohang has the optimal requirements for an AI computing center to be located, including abundant high-tech research facilities, talent, and stable power supply and demand.First of all, 4th generation radiation accelerators, POSTECH AI Research Institute, Apple R&D Support Center, and Robot Convergence Research Institute are integrated, making it easy to develop related technologies through research activities and having an environment where large-scale data processing and analysis are essential for the development of AI technology. In addition, there are plenty of master's and doctoral-level researchers in the AI field, centering on POSTECH and Handong University.As nuclear power plants are concentrated nearby, it is possible to supply stable power, which is important for fostering the AI industry, and if new and renewable energy generation facilities such as hydrogen fuel cells and offshore wind power are expanded, it is possible to supply low-cost power using the Special Act on Promotion of Distributed Energy.In particular, Pohang, which is leading innovation in new industries such as secondary batteries, bio, and hydrogen along with the steel industry, a traditional manufacturing industry, has optimal conditions to build a smart manufacturing ecosystem that combines AI technology. In addition, AI computing-related infrastructure such as the 1.52 trillion won global data center campus development project, the establishment of an AI open innovation center, and the establishment of an AI accelerator center is also progressing smoothly.Pohang's development of the AI industry will lead to innovation in the entire industry through convergence with manufacturing, Korea's main industry, beyond just regional development. Pohang City intends to begin its journey to become a Sovereign (sovereign) AI hub city in Korea based on specific strategies, action plans, strong industrial base and human resources. I sincerely hope that the starting point will be the establishment of Pohang, an AI computing center, with 500,000 citizens.[Lee Kang Deok, Mayor of Pohang] | 5 |
사진 확대
Artificial intelligence (AI) technology is becoming a key driver of industries and economies in the 21st century, and competition for supremacy by countries around the world is intensifying day by day. In particular, not long ago, Chinese AI start-up "DeepSeek" launched a new AI model that is almost equivalent to the performance of its competitors at a much lower cost, shocking the United States and the rest of the world. The 'deep shock', which is recognized as a groundbreaking achievement that is called the 'Sputnik moment of AI', is recognized as a great threat to the United States, which has led the innovation of the AI industry.Amid intensifying competition for global hegemony in the AI industry, mainly in the United States and China, South Korea launched the National AI Committee headed by the president in September last year and entered full-fledged competition with the vision of becoming one of the "three AI powers." However, even in the AI industry, the gap between the metropolitan area and provinces is gradually widening. If this imbalance is not resolved, Korea will be eliminated from global competition.Pohang, which is transforming from a steel industry city into a high-tech new industrial city such as a secondary battery, has a wide range of conditions for the AI industry to grow rapidly. In terms of balanced regional development, it also fits the optimal location of the AI industry. This is why Pohang is actively seeking to attract a national AI computing center that the government is promoting. The National AI Computing Center, which will accelerate the leap forward as one of the three AI powerhouses, is a base infrastructure that combines advanced AI semiconductors and dedicated programs to produce useful results needed by companies and research institutes from vast amounts of data, and the government plans to build it in non-metropolitan areas. Pohang has the optimal requirements for an AI computing center to be located, including abundant high-tech research facilities, talent, and stable power supply and demand.First of all, 4th generation radiation accelerators, POSTECH AI Research Institute, Apple R&D Support Center, and Robot Convergence Research Institute are integrated, making it easy to develop related technologies through research activities and having an environment where large-scale data processing and analysis are essential for the development of AI technology. In addition, there are plenty of master's and doctoral-level researchers in the AI field, centering on POSTECH and Handong University.As nuclear power plants are concentrated nearby, it is possible to supply stable power, which is important for fostering the AI industry, and if new and renewable energy generation facilities such as hydrogen fuel cells and offshore wind power are expanded, it is possible to supply low-cost power using the Special Act on Promotion of Distributed Energy.In particular, Pohang, which is leading innovation in new industries such as secondary batteries, bio, and hydrogen along with the steel industry, a traditional manufacturing industry, has optimal conditions to build a smart manufacturing ecosystem that combines AI technology. In addition, AI computing-related infrastructure such as the 1.52 trillion won global data center campus development project, the establishment of an AI open innovation center, and the establishment of an AI accelerator center is also progressing smoothly.Pohang's development of the AI industry will lead to innovation in the entire industry through convergence with manufacturing, Korea's main industry, beyond just regional development. Pohang City intends to begin its journey to become a Sovereign (sovereign) AI hub city in Korea based on specific strategies, action plans, strong industrial base and human resources. I sincerely hope that the starting point will be the establishment of Pohang, an AI computing center, with 500,000 citizens.[Lee Kang Deok, Mayor of Pohang] | 5 |
사진 확대
Artificial intelligence (AI) technology is becoming a key driver of industries and economies in the 21st century, and competition for supremacy by countries around the world is intensifying day by day. In particular, not long ago, Chinese AI start-up "DeepSeek" launched a new AI model that is almost equivalent to the performance of its competitors at a much lower cost, shocking the United States and the rest of the world. The 'deep shock', which is recognized as a groundbreaking achievement that is called the 'Sputnik moment of AI', is recognized as a great threat to the United States, which has led the innovation of the AI industry.Amid intensifying competition for global hegemony in the AI industry, mainly in the United States and China, South Korea launched the National AI Committee headed by the president in September last year and entered full-fledged competition with the vision of becoming one of the "three AI powers." However, even in the AI industry, the gap between the metropolitan area and provinces is gradually widening. If this imbalance is not resolved, Korea will be eliminated from global competition.Pohang, which is transforming from a steel industry city into a high-tech new industrial city such as a secondary battery, has a wide range of conditions for the AI industry to grow rapidly. In terms of balanced regional development, it also fits the optimal location of the AI industry. This is why Pohang is actively seeking to attract a national AI computing center that the government is promoting. The National AI Computing Center, which will accelerate the leap forward as one of the three AI powerhouses, is a base infrastructure that combines advanced AI semiconductors and dedicated programs to produce useful results needed by companies and research institutes from vast amounts of data, and the government plans to build it in non-metropolitan areas. Pohang has the optimal requirements for an AI computing center to be located, including abundant high-tech research facilities, talent, and stable power supply and demand.First of all, 4th generation radiation accelerators, POSTECH AI Research Institute, Apple R&D Support Center, and Robot Convergence Research Institute are integrated, making it easy to develop related technologies through research activities and having an environment where large-scale data processing and analysis are essential for the development of AI technology. In addition, there are plenty of master's and doctoral-level researchers in the AI field, centering on POSTECH and Handong University.As nuclear power plants are concentrated nearby, it is possible to supply stable power, which is important for fostering the AI industry, and if new and renewable energy generation facilities such as hydrogen fuel cells and offshore wind power are expanded, it is possible to supply low-cost power using the Special Act on Promotion of Distributed Energy.In particular, Pohang, which is leading innovation in new industries such as secondary batteries, bio, and hydrogen along with the steel industry, a traditional manufacturing industry, has optimal conditions to build a smart manufacturing ecosystem that combines AI technology. In addition, AI computing-related infrastructure such as the 1.52 trillion won global data center campus development project, the establishment of an AI open innovation center, and the establishment of an AI accelerator center is also progressing smoothly.Pohang's development of the AI industry will lead to innovation in the entire industry through convergence with manufacturing, Korea's main industry, beyond just regional development. Pohang City intends to begin its journey to become a Sovereign (sovereign) AI hub city in Korea based on specific strategies, action plans, strong industrial base and human resources. I sincerely hope that the starting point will be the establishment of Pohang, an AI computing center, with 500,000 citizens.[Lee Kang Deok, Mayor of Pohang] | 5 |
KARACHI: A new artificial intelligence (AI) system jointly developed by researchers at New York University (NYU) and The Aga Khan University (AKU) in Karachi has exposed a stark disparity in urban green spaces across Pakistan’s largest city, highlighting critical vulnerabilities to climate change.
The study was led by Dr. Rumi Chunara, who serves as director of the NYU Center for Health Data Science and is a member of NYU Tandon’s Visualization Imaging and Data Analysis Center (VIDA), and included NYU’s Miao Zhang, Hajjra Arshad, Manzar Abbas, Hamzah Jahanzeb, Izza Tahir, Javerya Hassan and Dr. Zainab Samad from AKU. The researchers used advanced AI techniques to analyze satellite imagery and assess urban greenery in Karachi.
The research, published in the ACM Journal on Computing and Sustainable Societies, found that Karachi averages just 4.17 square meters of per capita green space, which is even less than one half of the World Health Organization’s (WHO) recommended nine square meters per person.
“It is the fifth most populous city in the world and in recent years has faced both deadly heat waves and urban flooding,” said Dr. Zainab Samad of The Aga Khan University.
“The greenspace availability varies significantly across union councils. Three union councils — Darsanno Channo, Murad Memon, and Gulshan-e-Hadeed — have the highest greenspace values of over 80 m² per capita, while five union councils — Darya Abad, Behar Colony, Chishti Nagar, Banaras Colony, and Gulshan Said — have the lowest values of less than 0.1 m² per capita.”
Dr. Samad said areas exceeding WHO’s green space recommendations were situated on Karachi’s periphery, particularly in the east.
The AI system, which achieved 89.4 percent accuracy and 90.6 percent reliability in identifying vegetation, represents a significant improvement over traditional satellite analysis, which typically achieves around 63 percent accuracy.
“To train the AI model, we create new images by shifting the hue of original satellite images,” Dr. Chunara told Arab News. “This technique helps the model better recognize diverse vegetation types.”
This “green augmentation” process enhances the model’s ability to distinguish trees from grass, even in complex urban environments, according to the expert.
The study also revealed a correlation between paved roads and increased green spaces, reflecting broader urban development patterns.
“The correlation reflects broader urban development patterns, where more developed areas with paved roads often have higher socioeconomic status, leading to better access to green spaces and urban infrastructure,” Dr. Chunara said.
The disparity in green space distribution has significant implications for public health and environmental sustainability, according to Dr. Samad. These benefits may be unequally distributed, with low-income areas often lacking vegetation that makes them hotter and more polluted.
“A combination of AI techniques could be used to not only identify deficiencies in green space but also be prescriptive about where greening could be most helpful and how this could be achieved,” she said.
The researchers emphasized the importance of making the AI system and its findings accessible to local authorities.
“Ensuring that the AI system and its findings are accessible and usable for local authorities in Pakistan is a crucial aspect of this research,” Dr. Chunara said. “We will also facilitate direct communication with local authorities to provide ongoing support and ensure that the data is effectively integrated into their planning processes.”
Policy recommendations derived from the research include prioritizing green spaces in urban planning, assessing areas where green space is most needed, and exploring potential locations for repurposing underutilized spaces for greenery.
“City planning can prioritize green spaces through master plans and zoning,” Dr. Samad said, adding that addressing the disparity in green space distribution requires interventions at various levels.
“Infrastructure initiatives like public parks and tree planting programs can enhance greenery, while community-based actions such as volunteer maintenance and tree adoption foster local involvement.”
Interestingly, the researchers compared Karachi to Singapore which, despite similar population density, provides 9.9 square meters of green space per person, exceeding the WHO target.
“In addition to Singapore, cities like Katmandu and Perth have implemented greening initiatives, such as the Green Katmandu Project and the Perth Urban Greening Strategy,” Dr. Chunara noted.
“Other cities, like Dubai, have integrated green space initiatives into their master plans to promote sustainable urban development.”
In Pakistan, however, a major challenge remains ensuring that local authorities can effectively use AI-driven research despite limited technical resources.
Dr. Chunara’s said her team was addressing this by creating accessible visualizations, data summaries and tailored reports.
“We are committed to making the findings actionable by creating clear visualizations, data summaries, and tailored reports that are easy to understand,” she said. “By engaging with authorities in a collaborative and user-friendly way, we aim to bridge the technical gap and empower them to make informed decisions for the city’s future.” | 5 |
KARACHI: A new artificial intelligence (AI) system jointly developed by researchers at New York University (NYU) and The Aga Khan University (AKU) in Karachi has exposed a stark disparity in urban green spaces across Pakistan’s largest city, highlighting critical vulnerabilities to climate change.
The study was led by Dr. Rumi Chunara, who serves as director of the NYU Center for Health Data Science and is a member of NYU Tandon’s Visualization Imaging and Data Analysis Center (VIDA), and included NYU’s Miao Zhang, Hajjra Arshad, Manzar Abbas, Hamzah Jahanzeb, Izza Tahir, Javerya Hassan and Dr. Zainab Samad from AKU. The researchers used advanced AI techniques to analyze satellite imagery and assess urban greenery in Karachi.
The research, published in the ACM Journal on Computing and Sustainable Societies, found that Karachi averages just 4.17 square meters of per capita green space, which is even less than one half of the World Health Organization’s (WHO) recommended nine square meters per person.
“It is the fifth most populous city in the world and in recent years has faced both deadly heat waves and urban flooding,” said Dr. Zainab Samad of The Aga Khan University.
“The greenspace availability varies significantly across union councils. Three union councils — Darsanno Channo, Murad Memon, and Gulshan-e-Hadeed — have the highest greenspace values of over 80 m² per capita, while five union councils — Darya Abad, Behar Colony, Chishti Nagar, Banaras Colony, and Gulshan Said — have the lowest values of less than 0.1 m² per capita.”
Dr. Samad said areas exceeding WHO’s green space recommendations were situated on Karachi’s periphery, particularly in the east.
The AI system, which achieved 89.4 percent accuracy and 90.6 percent reliability in identifying vegetation, represents a significant improvement over traditional satellite analysis, which typically achieves around 63 percent accuracy.
“To train the AI model, we create new images by shifting the hue of original satellite images,” Dr. Chunara told Arab News. “This technique helps the model better recognize diverse vegetation types.”
This “green augmentation” process enhances the model’s ability to distinguish trees from grass, even in complex urban environments, according to the expert.
The study also revealed a correlation between paved roads and increased green spaces, reflecting broader urban development patterns.
“The correlation reflects broader urban development patterns, where more developed areas with paved roads often have higher socioeconomic status, leading to better access to green spaces and urban infrastructure,” Dr. Chunara said.
The disparity in green space distribution has significant implications for public health and environmental sustainability, according to Dr. Samad. These benefits may be unequally distributed, with low-income areas often lacking vegetation that makes them hotter and more polluted.
“A combination of AI techniques could be used to not only identify deficiencies in green space but also be prescriptive about where greening could be most helpful and how this could be achieved,” she said.
The researchers emphasized the importance of making the AI system and its findings accessible to local authorities.
“Ensuring that the AI system and its findings are accessible and usable for local authorities in Pakistan is a crucial aspect of this research,” Dr. Chunara said. “We will also facilitate direct communication with local authorities to provide ongoing support and ensure that the data is effectively integrated into their planning processes.”
Policy recommendations derived from the research include prioritizing green spaces in urban planning, assessing areas where green space is most needed, and exploring potential locations for repurposing underutilized spaces for greenery.
“City planning can prioritize green spaces through master plans and zoning,” Dr. Samad said, adding that addressing the disparity in green space distribution requires interventions at various levels.
“Infrastructure initiatives like public parks and tree planting programs can enhance greenery, while community-based actions such as volunteer maintenance and tree adoption foster local involvement.”
Interestingly, the researchers compared Karachi to Singapore which, despite similar population density, provides 9.9 square meters of green space per person, exceeding the WHO target.
“In addition to Singapore, cities like Katmandu and Perth have implemented greening initiatives, such as the Green Katmandu Project and the Perth Urban Greening Strategy,” Dr. Chunara noted.
“Other cities, like Dubai, have integrated green space initiatives into their master plans to promote sustainable urban development.”
In Pakistan, however, a major challenge remains ensuring that local authorities can effectively use AI-driven research despite limited technical resources.
Dr. Chunara’s said her team was addressing this by creating accessible visualizations, data summaries and tailored reports.
“We are committed to making the findings actionable by creating clear visualizations, data summaries, and tailored reports that are easy to understand,” she said. “By engaging with authorities in a collaborative and user-friendly way, we aim to bridge the technical gap and empower them to make informed decisions for the city’s future.” | 5 |
KARACHI: A new artificial intelligence (AI) system jointly developed by researchers at New York University (NYU) and The Aga Khan University (AKU) in Karachi has exposed a stark disparity in urban green spaces across Pakistan’s largest city, highlighting critical vulnerabilities to climate change.
The study was led by Dr. Rumi Chunara, who serves as director of the NYU Center for Health Data Science and is a member of NYU Tandon’s Visualization Imaging and Data Analysis Center (VIDA), and included NYU’s Miao Zhang, Hajjra Arshad, Manzar Abbas, Hamzah Jahanzeb, Izza Tahir, Javerya Hassan and Dr. Zainab Samad from AKU. The researchers used advanced AI techniques to analyze satellite imagery and assess urban greenery in Karachi.
The research, published in the ACM Journal on Computing and Sustainable Societies, found that Karachi averages just 4.17 square meters of per capita green space, which is even less than one half of the World Health Organization’s (WHO) recommended nine square meters per person.
“It is the fifth most populous city in the world and in recent years has faced both deadly heat waves and urban flooding,” said Dr. Zainab Samad of The Aga Khan University.
“The greenspace availability varies significantly across union councils. Three union councils — Darsanno Channo, Murad Memon, and Gulshan-e-Hadeed — have the highest greenspace values of over 80 m² per capita, while five union councils — Darya Abad, Behar Colony, Chishti Nagar, Banaras Colony, and Gulshan Said — have the lowest values of less than 0.1 m² per capita.”
Dr. Samad said areas exceeding WHO’s green space recommendations were situated on Karachi’s periphery, particularly in the east.
The AI system, which achieved 89.4 percent accuracy and 90.6 percent reliability in identifying vegetation, represents a significant improvement over traditional satellite analysis, which typically achieves around 63 percent accuracy.
“To train the AI model, we create new images by shifting the hue of original satellite images,” Dr. Chunara told Arab News. “This technique helps the model better recognize diverse vegetation types.”
This “green augmentation” process enhances the model’s ability to distinguish trees from grass, even in complex urban environments, according to the expert.
The study also revealed a correlation between paved roads and increased green spaces, reflecting broader urban development patterns.
“The correlation reflects broader urban development patterns, where more developed areas with paved roads often have higher socioeconomic status, leading to better access to green spaces and urban infrastructure,” Dr. Chunara said.
The disparity in green space distribution has significant implications for public health and environmental sustainability, according to Dr. Samad. These benefits may be unequally distributed, with low-income areas often lacking vegetation that makes them hotter and more polluted.
“A combination of AI techniques could be used to not only identify deficiencies in green space but also be prescriptive about where greening could be most helpful and how this could be achieved,” she said.
The researchers emphasized the importance of making the AI system and its findings accessible to local authorities.
“Ensuring that the AI system and its findings are accessible and usable for local authorities in Pakistan is a crucial aspect of this research,” Dr. Chunara said. “We will also facilitate direct communication with local authorities to provide ongoing support and ensure that the data is effectively integrated into their planning processes.”
Policy recommendations derived from the research include prioritizing green spaces in urban planning, assessing areas where green space is most needed, and exploring potential locations for repurposing underutilized spaces for greenery.
“City planning can prioritize green spaces through master plans and zoning,” Dr. Samad said, adding that addressing the disparity in green space distribution requires interventions at various levels.
“Infrastructure initiatives like public parks and tree planting programs can enhance greenery, while community-based actions such as volunteer maintenance and tree adoption foster local involvement.”
Interestingly, the researchers compared Karachi to Singapore which, despite similar population density, provides 9.9 square meters of green space per person, exceeding the WHO target.
“In addition to Singapore, cities like Katmandu and Perth have implemented greening initiatives, such as the Green Katmandu Project and the Perth Urban Greening Strategy,” Dr. Chunara noted.
“Other cities, like Dubai, have integrated green space initiatives into their master plans to promote sustainable urban development.”
In Pakistan, however, a major challenge remains ensuring that local authorities can effectively use AI-driven research despite limited technical resources.
Dr. Chunara’s said her team was addressing this by creating accessible visualizations, data summaries and tailored reports.
“We are committed to making the findings actionable by creating clear visualizations, data summaries, and tailored reports that are easy to understand,” she said. “By engaging with authorities in a collaborative and user-friendly way, we aim to bridge the technical gap and empower them to make informed decisions for the city’s future.” | 5 |
KARACHI: A new artificial intelligence (AI) system jointly developed by researchers at New York University (NYU) and The Aga Khan University (AKU) in Karachi has exposed a stark disparity in urban green spaces across Pakistan’s largest city, highlighting critical vulnerabilities to climate change.
The study was led by Dr. Rumi Chunara, who serves as director of the NYU Center for Health Data Science and is a member of NYU Tandon’s Visualization Imaging and Data Analysis Center (VIDA), and included NYU’s Miao Zhang, Hajjra Arshad, Manzar Abbas, Hamzah Jahanzeb, Izza Tahir, Javerya Hassan and Dr. Zainab Samad from AKU. The researchers used advanced AI techniques to analyze satellite imagery and assess urban greenery in Karachi.
The research, published in the ACM Journal on Computing and Sustainable Societies, found that Karachi averages just 4.17 square meters of per capita green space, which is even less than one half of the World Health Organization’s (WHO) recommended nine square meters per person.
“It is the fifth most populous city in the world and in recent years has faced both deadly heat waves and urban flooding,” said Dr. Zainab Samad of The Aga Khan University.
“The greenspace availability varies significantly across union councils. Three union councils — Darsanno Channo, Murad Memon, and Gulshan-e-Hadeed — have the highest greenspace values of over 80 m² per capita, while five union councils — Darya Abad, Behar Colony, Chishti Nagar, Banaras Colony, and Gulshan Said — have the lowest values of less than 0.1 m² per capita.”
Dr. Samad said areas exceeding WHO’s green space recommendations were situated on Karachi’s periphery, particularly in the east.
The AI system, which achieved 89.4 percent accuracy and 90.6 percent reliability in identifying vegetation, represents a significant improvement over traditional satellite analysis, which typically achieves around 63 percent accuracy.
“To train the AI model, we create new images by shifting the hue of original satellite images,” Dr. Chunara told Arab News. “This technique helps the model better recognize diverse vegetation types.”
This “green augmentation” process enhances the model’s ability to distinguish trees from grass, even in complex urban environments, according to the expert.
The study also revealed a correlation between paved roads and increased green spaces, reflecting broader urban development patterns.
“The correlation reflects broader urban development patterns, where more developed areas with paved roads often have higher socioeconomic status, leading to better access to green spaces and urban infrastructure,” Dr. Chunara said.
The disparity in green space distribution has significant implications for public health and environmental sustainability, according to Dr. Samad. These benefits may be unequally distributed, with low-income areas often lacking vegetation that makes them hotter and more polluted.
“A combination of AI techniques could be used to not only identify deficiencies in green space but also be prescriptive about where greening could be most helpful and how this could be achieved,” she said.
The researchers emphasized the importance of making the AI system and its findings accessible to local authorities.
“Ensuring that the AI system and its findings are accessible and usable for local authorities in Pakistan is a crucial aspect of this research,” Dr. Chunara said. “We will also facilitate direct communication with local authorities to provide ongoing support and ensure that the data is effectively integrated into their planning processes.”
Policy recommendations derived from the research include prioritizing green spaces in urban planning, assessing areas where green space is most needed, and exploring potential locations for repurposing underutilized spaces for greenery.
“City planning can prioritize green spaces through master plans and zoning,” Dr. Samad said, adding that addressing the disparity in green space distribution requires interventions at various levels.
“Infrastructure initiatives like public parks and tree planting programs can enhance greenery, while community-based actions such as volunteer maintenance and tree adoption foster local involvement.”
Interestingly, the researchers compared Karachi to Singapore which, despite similar population density, provides 9.9 square meters of green space per person, exceeding the WHO target.
“In addition to Singapore, cities like Katmandu and Perth have implemented greening initiatives, such as the Green Katmandu Project and the Perth Urban Greening Strategy,” Dr. Chunara noted.
“Other cities, like Dubai, have integrated green space initiatives into their master plans to promote sustainable urban development.”
In Pakistan, however, a major challenge remains ensuring that local authorities can effectively use AI-driven research despite limited technical resources.
Dr. Chunara’s said her team was addressing this by creating accessible visualizations, data summaries and tailored reports.
“We are committed to making the findings actionable by creating clear visualizations, data summaries, and tailored reports that are easy to understand,” she said. “By engaging with authorities in a collaborative and user-friendly way, we aim to bridge the technical gap and empower them to make informed decisions for the city’s future.” | 5 |
KARACHI: A new artificial intelligence (AI) system jointly developed by researchers at New York University (NYU) and The Aga Khan University (AKU) in Karachi has exposed a stark disparity in urban green spaces across Pakistan’s largest city, highlighting critical vulnerabilities to climate change.
The study was led by Dr. Rumi Chunara, who serves as director of the NYU Center for Health Data Science and is a member of NYU Tandon’s Visualization Imaging and Data Analysis Center (VIDA), and included NYU’s Miao Zhang, Hajjra Arshad, Manzar Abbas, Hamzah Jahanzeb, Izza Tahir, Javerya Hassan and Dr. Zainab Samad from AKU. The researchers used advanced AI techniques to analyze satellite imagery and assess urban greenery in Karachi.
The research, published in the ACM Journal on Computing and Sustainable Societies, found that Karachi averages just 4.17 square meters of per capita green space, which is even less than one half of the World Health Organization’s (WHO) recommended nine square meters per person.
“It is the fifth most populous city in the world and in recent years has faced both deadly heat waves and urban flooding,” said Dr. Zainab Samad of The Aga Khan University.
“The greenspace availability varies significantly across union councils. Three union councils — Darsanno Channo, Murad Memon, and Gulshan-e-Hadeed — have the highest greenspace values of over 80 m² per capita, while five union councils — Darya Abad, Behar Colony, Chishti Nagar, Banaras Colony, and Gulshan Said — have the lowest values of less than 0.1 m² per capita.”
Dr. Samad said areas exceeding WHO’s green space recommendations were situated on Karachi’s periphery, particularly in the east.
The AI system, which achieved 89.4 percent accuracy and 90.6 percent reliability in identifying vegetation, represents a significant improvement over traditional satellite analysis, which typically achieves around 63 percent accuracy.
“To train the AI model, we create new images by shifting the hue of original satellite images,” Dr. Chunara told Arab News. “This technique helps the model better recognize diverse vegetation types.”
This “green augmentation” process enhances the model’s ability to distinguish trees from grass, even in complex urban environments, according to the expert.
The study also revealed a correlation between paved roads and increased green spaces, reflecting broader urban development patterns.
“The correlation reflects broader urban development patterns, where more developed areas with paved roads often have higher socioeconomic status, leading to better access to green spaces and urban infrastructure,” Dr. Chunara said.
The disparity in green space distribution has significant implications for public health and environmental sustainability, according to Dr. Samad. These benefits may be unequally distributed, with low-income areas often lacking vegetation that makes them hotter and more polluted.
“A combination of AI techniques could be used to not only identify deficiencies in green space but also be prescriptive about where greening could be most helpful and how this could be achieved,” she said.
The researchers emphasized the importance of making the AI system and its findings accessible to local authorities.
“Ensuring that the AI system and its findings are accessible and usable for local authorities in Pakistan is a crucial aspect of this research,” Dr. Chunara said. “We will also facilitate direct communication with local authorities to provide ongoing support and ensure that the data is effectively integrated into their planning processes.”
Policy recommendations derived from the research include prioritizing green spaces in urban planning, assessing areas where green space is most needed, and exploring potential locations for repurposing underutilized spaces for greenery.
“City planning can prioritize green spaces through master plans and zoning,” Dr. Samad said, adding that addressing the disparity in green space distribution requires interventions at various levels.
“Infrastructure initiatives like public parks and tree planting programs can enhance greenery, while community-based actions such as volunteer maintenance and tree adoption foster local involvement.”
Interestingly, the researchers compared Karachi to Singapore which, despite similar population density, provides 9.9 square meters of green space per person, exceeding the WHO target.
“In addition to Singapore, cities like Katmandu and Perth have implemented greening initiatives, such as the Green Katmandu Project and the Perth Urban Greening Strategy,” Dr. Chunara noted.
“Other cities, like Dubai, have integrated green space initiatives into their master plans to promote sustainable urban development.”
In Pakistan, however, a major challenge remains ensuring that local authorities can effectively use AI-driven research despite limited technical resources.
Dr. Chunara’s said her team was addressing this by creating accessible visualizations, data summaries and tailored reports.
“We are committed to making the findings actionable by creating clear visualizations, data summaries, and tailored reports that are easy to understand,” she said. “By engaging with authorities in a collaborative and user-friendly way, we aim to bridge the technical gap and empower them to make informed decisions for the city’s future.” | 5 |
The Ministry of Education has selected four premier institutes to spearhead India’s AI Centers of Excellence (CoEs). AIIMS Delhi, IIT Delhi, IIT Kanpur and IIT Ropar have been chosen to lead the CoEs in AI with an emphasis on health, sustainable cities and agriculture, respectively, agriculture minister Shivraj Singh Chouhan informed Parliament recently. By bringing together premier research institutions, industry leaders, and government bodies, the CoEs aim to bridge the gap between academia and industry, ensuring that research is not just theoretical but translates into real-world solutions. Seamless industry integration, availability of high-quality annotated datasets, scalability of AI solutions for diverse Indian contexts, and regulatory concerns surrounding AI ethics and data privacy are some of the issues that the CoE will address. Ensuring that research does not remain confined to academic outputs but translates into real-world impact is one of the key hurdles. Additionally, the AI sector in India is still evolving, and the lack of structured AI datasets in areas like agriculture and healthcare poses a challenge for model training and deployment. Another critical issue is ensuring that AI solutions are accessible to all, including small-scale farmers and urban planners in tier-2 and tier-3 cities. “The focus areas of the CoEs include disease detection and healthcare advancements, precision farming techniques for improved agricultural productivity, and AI-powered smart city solutions for urban sustainability. By creating scalable AI models, the initiative seeks to integrate these technologies with national programmes, ensuring large-scale deployment and impact,” says Prof Manoj Gaur, director, IIT Jammu and Principal Investigator (PI) of this initiative which has been allocated a total financial outlay of Rs 990 crore over a period starting from 2023-24 to 2027-28 as part of the government's ‘Make AI in India and make AI work for India’ vision.
Since the initiative directly supports the ‘Make AI in India’ vision, it will focus on the development of indigenous AI technologies rather than relying on foreign solutions. “Additionally, part of the funding will be used to create open-source AI tools and datasets, particularly in the agriculture sector, to ensure that small farmers and rural enterprises can benefit from AI without high adoption costs,” says Rajeev Ahuja, director, IIT Ropar pointing to its engagement with AI-CoE in Agriculture which is focused on providing AI solutions to enhance precision farming, crop health monitoring, post-harvest technologies, to name a few.
Differentiating features
Unlike conventional academic research centres in Higher Educational Institutions (HEIs) which primarily focus on fundamental research and theoretical advancements, the AI CoEs are designed to be industry-oriented and multidisciplinary in nature. “Their primary mandate is to develop AI applications that are commercially viable, ensuring that the research conducted does not remain confined to academic papers but leads to tangible technological solutions. Another key distinction is the direct integration of these CoEs with industries and startups, facilitating innovation in collaboration with practitioners who understand market demands. Unlike academic research centres that function independently within their respective institutions, these CoEs operate as part of a national AI ecosystem, bringing together multiple institutions, including IITs, NITs, IIITs, and other stakeholders, to work on unified research objectives,” Prof Ahuja says.
Through industry-driven projects, internships, and mentorship programmes within academic settings, the CoEs will ensure that graduates gain hands-on experience with AI technologies. “Moreover, the CoEs will substantially aid in incorporating AI into courses by introducing emerging AI trends, interdisciplinary learning, and industry-relevant case studies. Importantly, the aim is that these three CoEs will work closely with the newly announced AI CoE for Education to potentially influence the educational curriculum,” Prof Gaur says.
Discussing the initiative’s broader impact, Prof Rangan Banerjee, director, IIT Delhi elaborates, “Each of the CoEs set up as independent section 8 not-for-profit companies, are a part of the government’s mandate in the Union Budget for AI integration in various fields with the objective of making a difference in people’s lives. IIT Delhi, for instance, is jointly engaged with AIIMS Delhi for one of the CoEs in the healthcare domain. Our aim is to catalyse research and look at ways to use AI for enabling more cost-effective healthcare delivery through a host of projects which will involve industry engagement and PhD scholars.”
Highlighting that the initiative is still in its infancy, Banerjee says that each of the CoEs will typically need funding of around Rs 300 crore over a period of time to deliver on their proposed outcomes.
“In healthcare transformation, this grant would contribute towards manpower training - both from the healthcare and technical perspective–in addition to the development/utilisation of medical equipment, regulatory clearances, clinical trials and seed funding for companies formed out of this initiative,” says Dr Krithika Rangarajan, associate professor, Oncoradiology, and PI, Centre of Excellence in Healthcare AI, AIIMS Delhi.
“There are several unresolved questions around safety, reliability, and liability related to AI technology. This is even more critical for our CoE in healthcare where there are immediate life and death questions involved,” says Chetan Arora, professor, Department of Computer Science and Engineering, IIT Delhi.
To maximise benefits, the challenges of the CoEs will be addressed through industry collaborations, government support, and open-data initiatives. Developing localised AI models that cater to regional requirements and language preferences will also ensure broader adoption.
Not exclusive to IITs
While the IITs serve as the lead institutions, the initiative is not exclusive to them. The selection process followed a rigorous nationwide call for proposals issued by the Ministry of Education (MoE) in November 2023. HEIs ranked within the top 100 in the NIRF Overall category were invited to apply. The selection of institutes was based on their strong research capabilities, expertise in AI, and extensive networks of academic and industry collaborations. “Each Centre functions as a nationwide consortium, bringing together top academic and research institutes, industry partners, startups, and other agencies. For instance, IIT Ropar’s Agriculture AI-CoE includes partners like IIT Guwahati, NIT Meghalaya, IIT Hyderabad, IIT Tirupati, and NIT Hamirpur amongst several others. The Health AI-CoE at AIIMS Delhi-IIT Delhi collaborates with institutions such as IIT Madras, IIT Bombay, AIIMS Patna, Ashoka University, MAHE-Manipal, IIIT Hyderabad, and IISc Bangalore. Similarly, the Sustainable Cities AI-CoE at IIT Kanpur involves key stakeholders, including Chennai Smart City, Bangalore Traffic Police, Adani Total Gas Ltd, IIT Hyderabad, IIT Gandhinagar, NIT Calicut, and IISc Bangalore,” Prof Gaur says. | 5 |
The Ministry of Education has selected four premier institutes to spearhead India’s AI Centers of Excellence (CoEs). AIIMS Delhi, IIT Delhi, IIT Kanpur and IIT Ropar have been chosen to lead the CoEs in AI with an emphasis on health, sustainable cities and agriculture, respectively, agriculture minister Shivraj Singh Chouhan informed Parliament recently. By bringing together premier research institutions, industry leaders, and government bodies, the CoEs aim to bridge the gap between academia and industry, ensuring that research is not just theoretical but translates into real-world solutions. Seamless industry integration, availability of high-quality annotated datasets, scalability of AI solutions for diverse Indian contexts, and regulatory concerns surrounding AI ethics and data privacy are some of the issues that the CoE will address. Ensuring that research does not remain confined to academic outputs but translates into real-world impact is one of the key hurdles. Additionally, the AI sector in India is still evolving, and the lack of structured AI datasets in areas like agriculture and healthcare poses a challenge for model training and deployment. Another critical issue is ensuring that AI solutions are accessible to all, including small-scale farmers and urban planners in tier-2 and tier-3 cities. “The focus areas of the CoEs include disease detection and healthcare advancements, precision farming techniques for improved agricultural productivity, and AI-powered smart city solutions for urban sustainability. By creating scalable AI models, the initiative seeks to integrate these technologies with national programmes, ensuring large-scale deployment and impact,” says Prof Manoj Gaur, director, IIT Jammu and Principal Investigator (PI) of this initiative which has been allocated a total financial outlay of Rs 990 crore over a period starting from 2023-24 to 2027-28 as part of the government's ‘Make AI in India and make AI work for India’ vision.
Since the initiative directly supports the ‘Make AI in India’ vision, it will focus on the development of indigenous AI technologies rather than relying on foreign solutions. “Additionally, part of the funding will be used to create open-source AI tools and datasets, particularly in the agriculture sector, to ensure that small farmers and rural enterprises can benefit from AI without high adoption costs,” says Rajeev Ahuja, director, IIT Ropar pointing to its engagement with AI-CoE in Agriculture which is focused on providing AI solutions to enhance precision farming, crop health monitoring, post-harvest technologies, to name a few.
Differentiating features
Unlike conventional academic research centres in Higher Educational Institutions (HEIs) which primarily focus on fundamental research and theoretical advancements, the AI CoEs are designed to be industry-oriented and multidisciplinary in nature. “Their primary mandate is to develop AI applications that are commercially viable, ensuring that the research conducted does not remain confined to academic papers but leads to tangible technological solutions. Another key distinction is the direct integration of these CoEs with industries and startups, facilitating innovation in collaboration with practitioners who understand market demands. Unlike academic research centres that function independently within their respective institutions, these CoEs operate as part of a national AI ecosystem, bringing together multiple institutions, including IITs, NITs, IIITs, and other stakeholders, to work on unified research objectives,” Prof Ahuja says.
Through industry-driven projects, internships, and mentorship programmes within academic settings, the CoEs will ensure that graduates gain hands-on experience with AI technologies. “Moreover, the CoEs will substantially aid in incorporating AI into courses by introducing emerging AI trends, interdisciplinary learning, and industry-relevant case studies. Importantly, the aim is that these three CoEs will work closely with the newly announced AI CoE for Education to potentially influence the educational curriculum,” Prof Gaur says.
Discussing the initiative’s broader impact, Prof Rangan Banerjee, director, IIT Delhi elaborates, “Each of the CoEs set up as independent section 8 not-for-profit companies, are a part of the government’s mandate in the Union Budget for AI integration in various fields with the objective of making a difference in people’s lives. IIT Delhi, for instance, is jointly engaged with AIIMS Delhi for one of the CoEs in the healthcare domain. Our aim is to catalyse research and look at ways to use AI for enabling more cost-effective healthcare delivery through a host of projects which will involve industry engagement and PhD scholars.”
Highlighting that the initiative is still in its infancy, Banerjee says that each of the CoEs will typically need funding of around Rs 300 crore over a period of time to deliver on their proposed outcomes.
“In healthcare transformation, this grant would contribute towards manpower training - both from the healthcare and technical perspective–in addition to the development/utilisation of medical equipment, regulatory clearances, clinical trials and seed funding for companies formed out of this initiative,” says Dr Krithika Rangarajan, associate professor, Oncoradiology, and PI, Centre of Excellence in Healthcare AI, AIIMS Delhi.
“There are several unresolved questions around safety, reliability, and liability related to AI technology. This is even more critical for our CoE in healthcare where there are immediate life and death questions involved,” says Chetan Arora, professor, Department of Computer Science and Engineering, IIT Delhi.
To maximise benefits, the challenges of the CoEs will be addressed through industry collaborations, government support, and open-data initiatives. Developing localised AI models that cater to regional requirements and language preferences will also ensure broader adoption.
Not exclusive to IITs
While the IITs serve as the lead institutions, the initiative is not exclusive to them. The selection process followed a rigorous nationwide call for proposals issued by the Ministry of Education (MoE) in November 2023. HEIs ranked within the top 100 in the NIRF Overall category were invited to apply. The selection of institutes was based on their strong research capabilities, expertise in AI, and extensive networks of academic and industry collaborations. “Each Centre functions as a nationwide consortium, bringing together top academic and research institutes, industry partners, startups, and other agencies. For instance, IIT Ropar’s Agriculture AI-CoE includes partners like IIT Guwahati, NIT Meghalaya, IIT Hyderabad, IIT Tirupati, and NIT Hamirpur amongst several others. The Health AI-CoE at AIIMS Delhi-IIT Delhi collaborates with institutions such as IIT Madras, IIT Bombay, AIIMS Patna, Ashoka University, MAHE-Manipal, IIIT Hyderabad, and IISc Bangalore. Similarly, the Sustainable Cities AI-CoE at IIT Kanpur involves key stakeholders, including Chennai Smart City, Bangalore Traffic Police, Adani Total Gas Ltd, IIT Hyderabad, IIT Gandhinagar, NIT Calicut, and IISc Bangalore,” Prof Gaur says. | 5 |
The Ministry of Education has selected four premier institutes to spearhead India’s AI Centers of Excellence (CoEs). AIIMS Delhi, IIT Delhi, IIT Kanpur and IIT Ropar have been chosen to lead the CoEs in AI with an emphasis on health, sustainable cities and agriculture, respectively, agriculture minister Shivraj Singh Chouhan informed Parliament recently. By bringing together premier research institutions, industry leaders, and government bodies, the CoEs aim to bridge the gap between academia and industry, ensuring that research is not just theoretical but translates into real-world solutions. Seamless industry integration, availability of high-quality annotated datasets, scalability of AI solutions for diverse Indian contexts, and regulatory concerns surrounding AI ethics and data privacy are some of the issues that the CoE will address. Ensuring that research does not remain confined to academic outputs but translates into real-world impact is one of the key hurdles. Additionally, the AI sector in India is still evolving, and the lack of structured AI datasets in areas like agriculture and healthcare poses a challenge for model training and deployment. Another critical issue is ensuring that AI solutions are accessible to all, including small-scale farmers and urban planners in tier-2 and tier-3 cities. “The focus areas of the CoEs include disease detection and healthcare advancements, precision farming techniques for improved agricultural productivity, and AI-powered smart city solutions for urban sustainability. By creating scalable AI models, the initiative seeks to integrate these technologies with national programmes, ensuring large-scale deployment and impact,” says Prof Manoj Gaur, director, IIT Jammu and Principal Investigator (PI) of this initiative which has been allocated a total financial outlay of Rs 990 crore over a period starting from 2023-24 to 2027-28 as part of the government's ‘Make AI in India and make AI work for India’ vision.
Since the initiative directly supports the ‘Make AI in India’ vision, it will focus on the development of indigenous AI technologies rather than relying on foreign solutions. “Additionally, part of the funding will be used to create open-source AI tools and datasets, particularly in the agriculture sector, to ensure that small farmers and rural enterprises can benefit from AI without high adoption costs,” says Rajeev Ahuja, director, IIT Ropar pointing to its engagement with AI-CoE in Agriculture which is focused on providing AI solutions to enhance precision farming, crop health monitoring, post-harvest technologies, to name a few.
Differentiating features
Unlike conventional academic research centres in Higher Educational Institutions (HEIs) which primarily focus on fundamental research and theoretical advancements, the AI CoEs are designed to be industry-oriented and multidisciplinary in nature. “Their primary mandate is to develop AI applications that are commercially viable, ensuring that the research conducted does not remain confined to academic papers but leads to tangible technological solutions. Another key distinction is the direct integration of these CoEs with industries and startups, facilitating innovation in collaboration with practitioners who understand market demands. Unlike academic research centres that function independently within their respective institutions, these CoEs operate as part of a national AI ecosystem, bringing together multiple institutions, including IITs, NITs, IIITs, and other stakeholders, to work on unified research objectives,” Prof Ahuja says.
Through industry-driven projects, internships, and mentorship programmes within academic settings, the CoEs will ensure that graduates gain hands-on experience with AI technologies. “Moreover, the CoEs will substantially aid in incorporating AI into courses by introducing emerging AI trends, interdisciplinary learning, and industry-relevant case studies. Importantly, the aim is that these three CoEs will work closely with the newly announced AI CoE for Education to potentially influence the educational curriculum,” Prof Gaur says.
Discussing the initiative’s broader impact, Prof Rangan Banerjee, director, IIT Delhi elaborates, “Each of the CoEs set up as independent section 8 not-for-profit companies, are a part of the government’s mandate in the Union Budget for AI integration in various fields with the objective of making a difference in people’s lives. IIT Delhi, for instance, is jointly engaged with AIIMS Delhi for one of the CoEs in the healthcare domain. Our aim is to catalyse research and look at ways to use AI for enabling more cost-effective healthcare delivery through a host of projects which will involve industry engagement and PhD scholars.”
Highlighting that the initiative is still in its infancy, Banerjee says that each of the CoEs will typically need funding of around Rs 300 crore over a period of time to deliver on their proposed outcomes.
“In healthcare transformation, this grant would contribute towards manpower training - both from the healthcare and technical perspective–in addition to the development/utilisation of medical equipment, regulatory clearances, clinical trials and seed funding for companies formed out of this initiative,” says Dr Krithika Rangarajan, associate professor, Oncoradiology, and PI, Centre of Excellence in Healthcare AI, AIIMS Delhi.
“There are several unresolved questions around safety, reliability, and liability related to AI technology. This is even more critical for our CoE in healthcare where there are immediate life and death questions involved,” says Chetan Arora, professor, Department of Computer Science and Engineering, IIT Delhi.
To maximise benefits, the challenges of the CoEs will be addressed through industry collaborations, government support, and open-data initiatives. Developing localised AI models that cater to regional requirements and language preferences will also ensure broader adoption.
Not exclusive to IITs
While the IITs serve as the lead institutions, the initiative is not exclusive to them. The selection process followed a rigorous nationwide call for proposals issued by the Ministry of Education (MoE) in November 2023. HEIs ranked within the top 100 in the NIRF Overall category were invited to apply. The selection of institutes was based on their strong research capabilities, expertise in AI, and extensive networks of academic and industry collaborations. “Each Centre functions as a nationwide consortium, bringing together top academic and research institutes, industry partners, startups, and other agencies. For instance, IIT Ropar’s Agriculture AI-CoE includes partners like IIT Guwahati, NIT Meghalaya, IIT Hyderabad, IIT Tirupati, and NIT Hamirpur amongst several others. The Health AI-CoE at AIIMS Delhi-IIT Delhi collaborates with institutions such as IIT Madras, IIT Bombay, AIIMS Patna, Ashoka University, MAHE-Manipal, IIIT Hyderabad, and IISc Bangalore. Similarly, the Sustainable Cities AI-CoE at IIT Kanpur involves key stakeholders, including Chennai Smart City, Bangalore Traffic Police, Adani Total Gas Ltd, IIT Hyderabad, IIT Gandhinagar, NIT Calicut, and IISc Bangalore,” Prof Gaur says. | 5 |
The Ministry of Education has selected four premier institutes to spearhead India’s AI Centers of Excellence (CoEs). AIIMS Delhi, IIT Delhi, IIT Kanpur and IIT Ropar have been chosen to lead the CoEs in AI with an emphasis on health, sustainable cities and agriculture, respectively, agriculture minister Shivraj Singh Chouhan informed Parliament recently. By bringing together premier research institutions, industry leaders, and government bodies, the CoEs aim to bridge the gap between academia and industry, ensuring that research is not just theoretical but translates into real-world solutions. Seamless industry integration, availability of high-quality annotated datasets, scalability of AI solutions for diverse Indian contexts, and regulatory concerns surrounding AI ethics and data privacy are some of the issues that the CoE will address. Ensuring that research does not remain confined to academic outputs but translates into real-world impact is one of the key hurdles. Additionally, the AI sector in India is still evolving, and the lack of structured AI datasets in areas like agriculture and healthcare poses a challenge for model training and deployment. Another critical issue is ensuring that AI solutions are accessible to all, including small-scale farmers and urban planners in tier-2 and tier-3 cities. “The focus areas of the CoEs include disease detection and healthcare advancements, precision farming techniques for improved agricultural productivity, and AI-powered smart city solutions for urban sustainability. By creating scalable AI models, the initiative seeks to integrate these technologies with national programmes, ensuring large-scale deployment and impact,” says Prof Manoj Gaur, director, IIT Jammu and Principal Investigator (PI) of this initiative which has been allocated a total financial outlay of Rs 990 crore over a period starting from 2023-24 to 2027-28 as part of the government's ‘Make AI in India and make AI work for India’ vision.
Since the initiative directly supports the ‘Make AI in India’ vision, it will focus on the development of indigenous AI technologies rather than relying on foreign solutions. “Additionally, part of the funding will be used to create open-source AI tools and datasets, particularly in the agriculture sector, to ensure that small farmers and rural enterprises can benefit from AI without high adoption costs,” says Rajeev Ahuja, director, IIT Ropar pointing to its engagement with AI-CoE in Agriculture which is focused on providing AI solutions to enhance precision farming, crop health monitoring, post-harvest technologies, to name a few.
Differentiating features
Unlike conventional academic research centres in Higher Educational Institutions (HEIs) which primarily focus on fundamental research and theoretical advancements, the AI CoEs are designed to be industry-oriented and multidisciplinary in nature. “Their primary mandate is to develop AI applications that are commercially viable, ensuring that the research conducted does not remain confined to academic papers but leads to tangible technological solutions. Another key distinction is the direct integration of these CoEs with industries and startups, facilitating innovation in collaboration with practitioners who understand market demands. Unlike academic research centres that function independently within their respective institutions, these CoEs operate as part of a national AI ecosystem, bringing together multiple institutions, including IITs, NITs, IIITs, and other stakeholders, to work on unified research objectives,” Prof Ahuja says.
Through industry-driven projects, internships, and mentorship programmes within academic settings, the CoEs will ensure that graduates gain hands-on experience with AI technologies. “Moreover, the CoEs will substantially aid in incorporating AI into courses by introducing emerging AI trends, interdisciplinary learning, and industry-relevant case studies. Importantly, the aim is that these three CoEs will work closely with the newly announced AI CoE for Education to potentially influence the educational curriculum,” Prof Gaur says.
Discussing the initiative’s broader impact, Prof Rangan Banerjee, director, IIT Delhi elaborates, “Each of the CoEs set up as independent section 8 not-for-profit companies, are a part of the government’s mandate in the Union Budget for AI integration in various fields with the objective of making a difference in people’s lives. IIT Delhi, for instance, is jointly engaged with AIIMS Delhi for one of the CoEs in the healthcare domain. Our aim is to catalyse research and look at ways to use AI for enabling more cost-effective healthcare delivery through a host of projects which will involve industry engagement and PhD scholars.”
Highlighting that the initiative is still in its infancy, Banerjee says that each of the CoEs will typically need funding of around Rs 300 crore over a period of time to deliver on their proposed outcomes.
“In healthcare transformation, this grant would contribute towards manpower training - both from the healthcare and technical perspective–in addition to the development/utilisation of medical equipment, regulatory clearances, clinical trials and seed funding for companies formed out of this initiative,” says Dr Krithika Rangarajan, associate professor, Oncoradiology, and PI, Centre of Excellence in Healthcare AI, AIIMS Delhi.
“There are several unresolved questions around safety, reliability, and liability related to AI technology. This is even more critical for our CoE in healthcare where there are immediate life and death questions involved,” says Chetan Arora, professor, Department of Computer Science and Engineering, IIT Delhi.
To maximise benefits, the challenges of the CoEs will be addressed through industry collaborations, government support, and open-data initiatives. Developing localised AI models that cater to regional requirements and language preferences will also ensure broader adoption.
Not exclusive to IITs
While the IITs serve as the lead institutions, the initiative is not exclusive to them. The selection process followed a rigorous nationwide call for proposals issued by the Ministry of Education (MoE) in November 2023. HEIs ranked within the top 100 in the NIRF Overall category were invited to apply. The selection of institutes was based on their strong research capabilities, expertise in AI, and extensive networks of academic and industry collaborations. “Each Centre functions as a nationwide consortium, bringing together top academic and research institutes, industry partners, startups, and other agencies. For instance, IIT Ropar’s Agriculture AI-CoE includes partners like IIT Guwahati, NIT Meghalaya, IIT Hyderabad, IIT Tirupati, and NIT Hamirpur amongst several others. The Health AI-CoE at AIIMS Delhi-IIT Delhi collaborates with institutions such as IIT Madras, IIT Bombay, AIIMS Patna, Ashoka University, MAHE-Manipal, IIIT Hyderabad, and IISc Bangalore. Similarly, the Sustainable Cities AI-CoE at IIT Kanpur involves key stakeholders, including Chennai Smart City, Bangalore Traffic Police, Adani Total Gas Ltd, IIT Hyderabad, IIT Gandhinagar, NIT Calicut, and IISc Bangalore,” Prof Gaur says. | 5 |
The Ministry of Education has selected four premier institutes to spearhead India’s AI Centers of Excellence (CoEs). AIIMS Delhi, IIT Delhi, IIT Kanpur and IIT Ropar have been chosen to lead the CoEs in AI with an emphasis on health, sustainable cities and agriculture, respectively, agriculture minister Shivraj Singh Chouhan informed Parliament recently. By bringing together premier research institutions, industry leaders, and government bodies, the CoEs aim to bridge the gap between academia and industry, ensuring that research is not just theoretical but translates into real-world solutions. Seamless industry integration, availability of high-quality annotated datasets, scalability of AI solutions for diverse Indian contexts, and regulatory concerns surrounding AI ethics and data privacy are some of the issues that the CoE will address. Ensuring that research does not remain confined to academic outputs but translates into real-world impact is one of the key hurdles. Additionally, the AI sector in India is still evolving, and the lack of structured AI datasets in areas like agriculture and healthcare poses a challenge for model training and deployment. Another critical issue is ensuring that AI solutions are accessible to all, including small-scale farmers and urban planners in tier-2 and tier-3 cities. “The focus areas of the CoEs include disease detection and healthcare advancements, precision farming techniques for improved agricultural productivity, and AI-powered smart city solutions for urban sustainability. By creating scalable AI models, the initiative seeks to integrate these technologies with national programmes, ensuring large-scale deployment and impact,” says Prof Manoj Gaur, director, IIT Jammu and Principal Investigator (PI) of this initiative which has been allocated a total financial outlay of Rs 990 crore over a period starting from 2023-24 to 2027-28 as part of the government's ‘Make AI in India and make AI work for India’ vision.
Since the initiative directly supports the ‘Make AI in India’ vision, it will focus on the development of indigenous AI technologies rather than relying on foreign solutions. “Additionally, part of the funding will be used to create open-source AI tools and datasets, particularly in the agriculture sector, to ensure that small farmers and rural enterprises can benefit from AI without high adoption costs,” says Rajeev Ahuja, director, IIT Ropar pointing to its engagement with AI-CoE in Agriculture which is focused on providing AI solutions to enhance precision farming, crop health monitoring, post-harvest technologies, to name a few.
Differentiating features
Unlike conventional academic research centres in Higher Educational Institutions (HEIs) which primarily focus on fundamental research and theoretical advancements, the AI CoEs are designed to be industry-oriented and multidisciplinary in nature. “Their primary mandate is to develop AI applications that are commercially viable, ensuring that the research conducted does not remain confined to academic papers but leads to tangible technological solutions. Another key distinction is the direct integration of these CoEs with industries and startups, facilitating innovation in collaboration with practitioners who understand market demands. Unlike academic research centres that function independently within their respective institutions, these CoEs operate as part of a national AI ecosystem, bringing together multiple institutions, including IITs, NITs, IIITs, and other stakeholders, to work on unified research objectives,” Prof Ahuja says.
Through industry-driven projects, internships, and mentorship programmes within academic settings, the CoEs will ensure that graduates gain hands-on experience with AI technologies. “Moreover, the CoEs will substantially aid in incorporating AI into courses by introducing emerging AI trends, interdisciplinary learning, and industry-relevant case studies. Importantly, the aim is that these three CoEs will work closely with the newly announced AI CoE for Education to potentially influence the educational curriculum,” Prof Gaur says.
Discussing the initiative’s broader impact, Prof Rangan Banerjee, director, IIT Delhi elaborates, “Each of the CoEs set up as independent section 8 not-for-profit companies, are a part of the government’s mandate in the Union Budget for AI integration in various fields with the objective of making a difference in people’s lives. IIT Delhi, for instance, is jointly engaged with AIIMS Delhi for one of the CoEs in the healthcare domain. Our aim is to catalyse research and look at ways to use AI for enabling more cost-effective healthcare delivery through a host of projects which will involve industry engagement and PhD scholars.”
Highlighting that the initiative is still in its infancy, Banerjee says that each of the CoEs will typically need funding of around Rs 300 crore over a period of time to deliver on their proposed outcomes.
“In healthcare transformation, this grant would contribute towards manpower training - both from the healthcare and technical perspective–in addition to the development/utilisation of medical equipment, regulatory clearances, clinical trials and seed funding for companies formed out of this initiative,” says Dr Krithika Rangarajan, associate professor, Oncoradiology, and PI, Centre of Excellence in Healthcare AI, AIIMS Delhi.
“There are several unresolved questions around safety, reliability, and liability related to AI technology. This is even more critical for our CoE in healthcare where there are immediate life and death questions involved,” says Chetan Arora, professor, Department of Computer Science and Engineering, IIT Delhi.
To maximise benefits, the challenges of the CoEs will be addressed through industry collaborations, government support, and open-data initiatives. Developing localised AI models that cater to regional requirements and language preferences will also ensure broader adoption.
Not exclusive to IITs
While the IITs serve as the lead institutions, the initiative is not exclusive to them. The selection process followed a rigorous nationwide call for proposals issued by the Ministry of Education (MoE) in November 2023. HEIs ranked within the top 100 in the NIRF Overall category were invited to apply. The selection of institutes was based on their strong research capabilities, expertise in AI, and extensive networks of academic and industry collaborations. “Each Centre functions as a nationwide consortium, bringing together top academic and research institutes, industry partners, startups, and other agencies. For instance, IIT Ropar’s Agriculture AI-CoE includes partners like IIT Guwahati, NIT Meghalaya, IIT Hyderabad, IIT Tirupati, and NIT Hamirpur amongst several others. The Health AI-CoE at AIIMS Delhi-IIT Delhi collaborates with institutions such as IIT Madras, IIT Bombay, AIIMS Patna, Ashoka University, MAHE-Manipal, IIIT Hyderabad, and IISc Bangalore. Similarly, the Sustainable Cities AI-CoE at IIT Kanpur involves key stakeholders, including Chennai Smart City, Bangalore Traffic Police, Adani Total Gas Ltd, IIT Hyderabad, IIT Gandhinagar, NIT Calicut, and IISc Bangalore,” Prof Gaur says. | 5 |
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We StartupNews.fyi want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected] | 5 |
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We StartupNews.fyi want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected] | 5 |
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We StartupNews.fyi want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected] | 5 |
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We StartupNews.fyi want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected] | 5 |
Disclaimer
We strive to uphold the highest ethical standards in all of our reporting and coverage. We StartupNews.fyi want to be transparent with our readers about any potential conflicts of interest that may arise in our work. It’s possible that some of the investors we feature may have connections to other businesses, including competitors or companies we write about. However, we want to assure our readers that this will not have any impact on the integrity or impartiality of our reporting. We are committed to delivering accurate, unbiased news and information to our audience, and we will continue to uphold our ethics and principles in all of our work. Thank you for your trust and support.
Website Upgradation is going on for any glitch kindly connect at [email protected] | 5 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.