Curiosity-driven research has long led to technological changes. A century ago, curiosity about atoms gave rise to quantum mechanics and eventually the transistor came to the center of modern computing. In contrast, the steam engine was a practical breakthrough, but it required fundamental research in thermodynamics to fully harness its power.
Today, artificial intelligence and science find themselves at a similar inflection point. The current AI revolution has been fueled by decades of research in the mathematical and physical sciences (MPS), which have provided challenging problems, datasets, and insights that have made modern AI possible. The 2024 Nobel Prizes in Physics and Chemistry, recognizing fundamental AI methods rooted in physics and AI applications to protein design, made this connection impossible to ignore.
In 2025, MIT hosts a Workshop on the Future of AI+MPSFunded by the National Science Foundation with support from the MIT School of Science and the MIT Departments of Physics, Chemistry, and Mathematics. The workshop brought together leading AI and science researchers to chart how MPS domains can best take advantage of – and contribute to – the future of AI. A white paper containing recommendations for funding agencies, institutions and researchers has now been released published in Machine Learning: Science and Technology. In this interview, Jesse Thaler, professor of physics at MIT and chair of the workshop, describes the key themes and how MIT is positioning itself to lead in AI and science.
Why: What are the main topics of the report regarding last year’s Assembly of Leaders in Mathematical and Physical Sciences?
A: It was exciting to gather so many researchers at the forefront of AI and science in one room. Although workshop participants came from five different scientific communities – astronomy, chemistry, materials science, mathematics and physics – we found many commonalities in the way they engaged with AI. A real consensus emerged from our animated discussions: coordinated investments in computing and data infrastructure, cross-disciplinary research techniques, and rigorous training can meaningfully advance both AI and science.
One of the key insights was that it should be a two-way street. It’s not just about using AI to improve science; Science can also improve AI. Scientists excel at gaining insights from complex systems, including neural networks, by uncovering underlying principles and emerging behaviors. We call this the “science of AI,” and it comes in three forms: the science that drives AI, where scientific reasoning informs fundamental AI approaches; Science Driven AI, where scientific challenges drive the development of new algorithms; and the science explaining AI, where scientific tools help explain how machine intelligence actually works.
For example, in my own field of particle physics, researchers are developing real-time AI algorithms to handle the data flood from collider experiments. This work has a direct impact on the discovery of new physics, but the algorithms themselves prove valuable far beyond our field. The workshop made clear that the science of AI should be a community priority – it has the potential to transform the way we understand, develop, and control AI systems.
Of course, connecting science and AI requires people who can work in both worlds. Attendees consistently stressed the need for “centaur scientists” – researchers with true interdisciplinary expertise. From integrated graduate courses to interdisciplinary PhD programs to joint faculty appointments, it emerged as essential to support these polymaths at every stage of their career.
Why: How do MIT’s AI and science efforts align with the workshop’s recommendations?
A: The workshop formulated its recommendations based on three pillars: research, talent and community. As director of the NSF Institute for Artificial Intelligence and Fundamental Interactions (IAFI) – a collaborative AI and physics effort between MIT and Harvard, Northeastern, and Tufts universities – I have seen firsthand how effective this framework can be. Extending this to MIT, we can see where progress is being made and where there are opportunities.
On the research front, MIT is already enabling AI-and-science work in both directions. Even a quick scroll MIT News It shows how individual researchers in the School of Science are leading AI-powered projects, building a pipeline of knowledge and uncovering new opportunities. Additionally, collaborative efforts such as IAIFI and the Accelerated AI Algorithms for Data-Driven Discovery (A3D3) Institute focus interdisciplinary energy for greater impact. The MIT Generative AI Impact Consortium is also supporting application-driven AI work at the university level.
To foster early-career AI-and-science talent, several initiatives are training the next generation of centaur scientists. The MIT Schwarzman College of Computing’s Common Ground for Computing Education program helps students become “bilingual” in computing and their home discipline. Interdisciplinary PhD pathways are also gaining popularity; IAIFI worked with the MIT Institute for Data, Systems, and Society to create an institute in physics, statistics, and data science, and about 10 percent of physics PhD students now choose it – a number that is likely to grow. Dedicated postdoctoral roles such as the IAIFI Fellowship and the Taiyabati Fellowship give early-career researchers the freedom to pursue interdisciplinary work. Funding Centaur scientists and giving them the space to make connections across domains, universities and career stages has been transformative.
Ultimately, community-building connects everyone together. Ranging from focused workshops to larger symposia, the organization of interdisciplinary events signals that AI and science is not a limited task – it is an emerging field. MIT has the talent and resources to make a significant impact, and hosting these celebrations at multiple levels helps establish that leadership.
Why: What lessons can MIT learn about how to further its AI-and-science efforts?
A: The workshop made some important points clear: The institutions that lead in AI and science will be those that think systematically, not piecemeal. Resources are limited, so priorities matter. Workshop attendees were clear about what is possible when an institution coordinates appointments, research and training based on a cohesive strategy.
MIT is well-positioned to build on what is already underway with more structural initiatives – joint faculty lines across computing and scientific domains, expanded interdisciplinary degree pathways, and intentional “science of AI” funding. We are already seeing steps in this direction; This year, the MIT Schwarzman College of Computing and the Physics departments are conducting their first joint faculty search, which is exciting to see.
The virtuous cycle of AI and science has the potential to be truly transformative – providing deeper insights into AI, accelerating scientific discovery, and creating robust tools for both. By developing an intentional strategy, MIT will be well-positioned to lead and benefit from the coming waves of AI.
