As artificial intelligence continues to redefine everything from manufacturing and marketing to academia and healthcare, concerns about its broader implications for knowledge creation, privacy threats, employment expectations,, and security risks are growing.
Now, a new cohort of investigators prepared to tackle these challenges has been named, with today’s announcement by Schmidt Sciences of its AI2050 Fellows for 2025. This year’s group consists of 28 recipients, including 7 senior scholars and 21 early career investigators.
Senior AI2050 Fellows are established leaders who’ve already made significant contributions to their field. Early Career Fellows are postdoctoral and pre-tenure researchers from around the world. All together they will receive more than $18 million in fellowship support for their work.
AI2050 is a project funded by Schmidt Sciences, a philanthropic initiative co-founded in 2024 by former Google CEO and Chairman Eric Schmidt and his wife Wendy, President of the Schmidt Family Foundation. It challenges researchers to imagine the year 2050, by which time AI has become extremely beneficial and to conduct research that addresses this defining question: “It’s 2050, and AI has turned out to be hugely beneficial to society — what happened?”
The project is aimed at advancing the capacity of artificial intelligence to tackle an evolving list of 10 “hard problems,” involving major scientific questions, technical issues, and risks associated with the revolutionary technology.
This year’s group is the 4th cohort of the program, which now numbers 99 fellows across eight countries and 42 institutions. “In four years, the AI2050 fellows have created a deep sense of community that we are grateful to be able to grow each year,” said Mark Greaves, executive director of AI2050, in a press release. “We trust that the network they’ve created will remain a source of inspiration and support throughout their careers as they advance AI for the benefit of all.”
Previous AI2050 fellows have helped make AI more trustworthy, improved the safety of self-driving cars, and improved chatbots for K-12 teachers. The fellows have also created datasets of cross-cultural human values and developed Humanity’s Last Exam, which tests human knowledge over a broad set of subjects, including mathematics, physics, biology/medicine, humanities/social sciences, computer science/artificial intelligence, engineering, chemistry, and other fields.
The work of the AI2050 fellows has been recognized with numerous awards and citations. Last year, they published more than 100 papers and posters in leading publications and scientific proceedings.
“AI is underhyped, especially when it comes to its potential to benefit humanity,” said Eric Schmidt, in the release, adding that "the AI2050 fellowship was established to turn that potential into reality—by supporting the people and ideas shaping a healthier, more resilient and more secure world.”
In addition to receiving financial support, AI2050 fellows participate in an annual gathering to share findings and learn from other experts in the field. They also can apply for additional funding to work with each other on AI2050-related projects.
The Senior Fellows
Alán Aspuru-Guzik (University of Toronto) aims to create a specialized AI chemist that will be able to collaborate with human chemists.
Surya Ganguli (Stanford University) will explore some of the first principles of neuroscience and AI, developing a scientific foundation for AI tools to be used in a trustworthy, safe and robust manner.
Shirley Ho (Flatiron Institute & New York University) intends to combine specialized scientific AI models like AlphaFold with more general AI models to advance scientific artificial general intelligence.
Sheila McIlraith (University of Toronto) seeks to equip AI with a purposeful theory of mind so that AI systems can understand beliefs, desires and intentions of themselves and others and choose actions that consider the welfare and agency of human beings.
Dawn Song (University of California at Berkeley) will develop AI tools that write computer code and offer mathematical proofs and security specifications demonstrating that the code is both correct and secure.
Philip Torr (University of Oxford) will developing an AI system capable of historical reasoning, interpreting evidence, testing theories of cultural and societal change, and exploring what makes societies more cohesive.
Luke Zettlemoyer (University of Washington) will challenge the assumption that AI models created by a central authority trained on all available data perform best. He aims to improve upon existing designs.
The 2025 AI2050 Early Career Fellows
● David Ifeoluwa Adelani, Assistant Professor, Mila - Quebec AI Institute
● Michael Albergo, Incoming Assistant Professor, Harvard University
● Antoine Bosselut, Assistant Professor, École Polytechnique Fédérale de Lausanne
● Charlotte Bunne, Assistant Professor, École Polytechnique Fédérale de Lausanne
● Miles Cranmer, Assistant Professor, University of Cambridge
● Tri Dao, Assistant Professor, Princeton University
● Adam Gleave, Co-founder and CEO, FAR.AI
● Surbhi Goel, Assistant Professor, University of Pennsylvania
● Brian Hie, Assistant Professor, Stanford University
● Natasha Jaques, Assistant Professor, University of Washington
● Hannah Kerner, Assistant Professor, Arizona State University
● Zongyi Li, Incoming Assistant Professor, New York University
● Krishna Pillutla, Assistant Professor, Indian Institute of Technology, Madras
● Lindsey Raymond, Incoming Assistant Professor, Massachusetts Institute of Technology
● Martin Schrimpf, Assistant Professor, École Polytechnique Fédérale de Lausanne
● Christian Schroeder de Witt, Research Fellow, University of Oxford
● Weiyan Shi, Assistant Professor, Northeastern University
● Tess Smidt, Associate Professor, Massachusetts Institute of Technology
● Melanie Weber, Assistant Professor, Harvard University
● Ellen Zhong, Assistant Professor, Princeton University
● Peichen Zhong, Assistant Professor, National University of Singapore
James Manyika, co-chair of AI2050 and a senior vice president at Google, praised the fellows as “ambitious yet collaborative researchers who focus on AI innovation and the opportunities and challenges in our AI2050 motivating question, and added, ”this technology can and will bring about an epochal shift in our society—and the AI2050 fellows are shaping that change so it is a benefit for all people.”