In a recent news interview, former Google CEO Eric Schmidt called AI, in no uncertain terms, “the most powerful technology that will be invented in our lifetimes.” It will come as no surprise that Schmidt, one of the most successful people in tech, has devoted much of his current philanthropy toward the development of technology — with artificial intelligence taking a central role.
We have been writing about Eric Schmidt’s hopes for AI and tech for several years. The Schmidts have been supporting AI research since before the 2024 formation of Schmidt Sciences, which “brings together the science-focused efforts of Eric and Wendy Schmidt.” The program I’m highlighting here is an academic fellowship to advance AI research, called AI2050, which IP’s Mike Scutari discussed back in 2022. The program comes at science funding in a unique way: It asks AI scholars to imagine the year 2050 and to (optimistically) envision some of the most beneficial AI-based solutions that could arise by that time. The nontraditional nature of the funding underscores Eric and Wendy Schmidt’s established pattern of backing promising and unproven but perhaps even risky avenues of research.
The AI2050 program, supported by a $125 million commitment over five years from the Schmidts, has been providing funding to help top established AI scholars, as well as up and comers, push their research forward. A few weeks ago, Schmidt Sciences named 25 AI scholars — five established leaders in the field and 20 early in their careers — to be the third and latest cohort of its AI2050 program.
The program provides a total of $12 million across the group of scholars to advance their research. They’re working in various disciplines and projects — some seem to be likely subjects for AI technology, such as drug discovery or particle physics, while others are perhaps more surprising, like the reduction of infant mortality. Including the previous fellows, AI2050 has now backed a total of 71 researchers from seven countries and dozens of institutions.
“The world has 8 billion people in it and we think AI is going to be a transformational technology for those 8 billion,” said Mark Greaves, executive director of the AI and Advanced Computing Institute at Schmidt Sciences. “Talent is not evenly distributed, and through the AI2050 program, we can bring up worldwide talent to solve problems, and not only look for talent where everyone else looks.”
AI2050 researchers tackle “hard problems” in AI
The program’s funding and fellowship decisions are guided in large part by what the Schmidt AI team calls a “list of hard problems in AI,” a set of general scientific and social concepts compiled by AI2050 co-chairs Eric Schmidt and James Manyika. The list, which is subject to change as the technology evolves, includes goals in science, technology and society — including comparatively obvious topics like health and life sciences, but also issues like economic scarcity and abundance, and many more arcane questions of how to improve the function of AI itself.
Senior AI2050 fellows from previous years include University of Cape Town researcher Kelly Chibale from the University of Cape Town; Fei-Fei Li, cofounder of World Labs; and Daniela Rus, who established Liquid AI and pioneered the use of liquid neural networks, which are smaller versions of AI neural networks that can continue to learn from stimuli after their initial training.
The newly named 2024 AI2050 senior fellows include economist David Autor, codirector of the Shaping the Future of Work Initiative at MIT, who investigates how people interact with new AI tools and how AI will shape employment and income in the future; Stanford AI researcher Yejin Choi, who develops methods to enable AI to reflect multiple human value systems; and Carla Gomes, director of the Institute for Computational Sustainability at Cornell University, who addresses limitations of AI for scientific discovery.
Related Inside Philanthropy Resources:
For Subscribers Only
No one can apply for an AI2050 fellowship — candidates are nominated by a group of nearly 300 AI experts around the world. “We’re driven by the people,” Greaves said. “The first bar is always scientific excellence, and then out of that pool, we can see how they lay across the hard problems.”
AI is one of the core focus areas of Schmidt Sciences, the science- and technology-oriented philanthropic organization Eric and Wendy Schmidt established about a year ago to take on the bulk of the couple’s many science and technology initiatives and interests. The science team that had been operating within Schmidt Futures shifted into the newer Schmidt Sciences organization.
Schmidt Sciences says its particular aim will be supporting researchers using technology to pursue early-stage, high-risk hypotheses spanning a range of disciplines, from basic science to pressing planetary problems. Also in the organization’s DNA is support for the increasingly interdisciplinary nature of science.
In addition to AI and advanced computing, Schmidt Sciences’ focus areas include astrophysics and space, biosciences, climate, and science systems. (The Schmidts’ longer-standing organizations, like Schmidt Futures and the Schmidt Family Foundation, continue to operate in support of a range of causes.)
Throughout this work, Schmidt Sciences — like many philanthropic organizations that have waded into the still fraught field of AI — stresses that emphasis will always be placed on safe and ethical uses of AI, and how the technology can be distributed so that people everywhere can use it and realize its benefits. This emphasis applies not just to the AI2050 program, but to other efforts within Schmidt Sciences’ AI focus, such as its new Humanities and AI Virtual Institute — which illustrates Schmidt philanthropy’s broad, optimistic conception of how the technology can be useful.
“AI has now reached a level of scale where philosophical questions and philosophical analyses can finally be relevant to the systems that the computer scientists built,” Greaves said. “Modern AI has shown itself to be unbelievably fluent in language, and understanding how language works,” he went on. “This interaction between AI computer science and humanities makes it a natural for understanding the kinds of materials which humanities scholars work with — language, art, media — so AI can be an incredible new tool to advance humanities scholarship.”
