The Simons Foundation, OpenAI and Anthropic have announced a $15 million initiative to help mathematicians navigate the growing influence of artificial intelligence on mathematical research. The programme will fund workshops, research initiatives and community-led efforts to evaluate AI-generated mathematical results and establish standards for using the technology.
Announced on October 15, 2026, the partnership comes as AI systems increasingly produce mathematical arguments, develop proofs and assist researchers with complex problems. While these capabilities offer new opportunities for discovery, they have also raised questions about verification, reliability and how mathematicians should assess a growing volume of AI-generated work.
Under the initiative, the three organisations will contribute a combined $15 million, with the Simons Foundation managing and distributing the funding through an open call for proposals. Anthropic and OpenAI will also provide AI tools and computing resources to researchers supported by the foundation.
How will the $15 million fund be used?
The initiative aims to give mathematicians greater resources to understand and evaluate AI-assisted research rather than simply encouraging the adoption of new tools.
The Simons Foundation plans to support workshops, conferences and other programmes where researchers can examine AI-generated results, discuss new research methods and share their experiences with emerging technologies.
The programme will also help the mathematical community develop shared standards for evaluating and using AI in research. This is particularly important as AI systems produce results at a pace that can make conventional human review challenging.
An open call for proposals will allow the community to help shape how the funding is allocated. The foundation has said that researchers, rather than any single technology company or institution, should drive discussions about how AI tools are assessed and adopted.
OpenAI and Anthropic to provide AI tools and computing resources
Beyond financial support, OpenAI and Anthropic will provide researchers backed by the Simons Foundation with access to AI tools and computing resources.
These resources are intended to help scientists investigate AI-assisted methods, explore mathematical problems and develop approaches for checking computer-generated results.
The collaboration will extend beyond mathematics to include basic science, where AI systems are increasingly being used to analyse data, generate hypotheses and assist with complex computational tasks.
Providing computing resources alongside funding could help researchers experiment with AI systems without having to build all the necessary infrastructure themselves. The initiative’s practical impact will depend on how those resources are distributed and whether researchers can use them to produce verifiable scientific advances.
Why are mathematicians reassessing AI’s role?
AI is becoming more capable of generating mathematical arguments and helping researchers explore problems that previously required extensive human effort.
However, producing a plausible-looking proof is not the same as establishing that a mathematical result is correct. AI-generated arguments can contain logical errors, omit necessary steps or present results without enough explanation for other researchers to evaluate them.
The volume of AI-generated work is another challenge. As systems produce more results, mathematicians must decide which findings deserve attention, which can be independently verified and which represent genuinely meaningful advances.
OpenAI has recently published mathematical results generated by one of its frontier AI models and said it wants to help the research community examine and build on that work. The new funding initiative provides a wider framework for researchers to evaluate such outputs collectively.
The effort therefore addresses not only the technical capabilities of AI but also the research practices needed to maintain confidence in mathematical discoveries.
Simons Foundation’s history of supporting mathematical research
The Simons Foundation has a long record of funding mathematics and basic science. Its earlier support for machine learning in scientific research includes backing the development of Lean, a proof assistant that allows mathematical arguments to be expressed in a form that software can check.
Lean helps researchers formalise proofs so that their logical steps can be verified by a computer. This does not determine whether a result is important or interesting, but it can provide strong assurance that a formalised proof follows the rules of the underlying mathematical system.
Tools of this kind are becoming increasingly relevant as AI systems generate more mathematical content. Researchers can use formal verification to check certain results rather than relying solely on manually reviewing lengthy arguments.
The foundation has also organised workshops, lectures and collaborations to bring mathematicians together to explore AI’s potential and limitations.
David Spergel calls for a collective approach to AI in mathematics
David Spergel, president of the Simons Foundation, said the initiative was intended to help researchers establish a shared understanding of how mathematics should evolve as AI capabilities advance.
“Now is the time to look at mathematics as a whole, not as a collection of proofs and theorems, but as a collective understanding driven by human curiosity,” Spergel said.
His remarks reflect the foundation’s emphasis on supporting collaboration across the mathematical community. The aim is to encourage researchers to shape how AI is used in their field instead of allowing individual companies or institutions to determine the direction alone.
The initiative also recognises that mathematical research involves more than proving statements correct. Researchers must identify worthwhile problems, interpret results and determine whether new findings meaningfully advance knowledge.
What the initiative could mean for the future of mathematics
The $15 million programme gives mathematicians an opportunity to examine how AI can support research while maintaining rigorous standards of verification.
Workshops and collaborative projects could help researchers share effective methods, develop clearer evaluation practices and improve the use of AI tools in both theoretical and computational mathematics.
The funding also comes at a time when the relationship between human researchers and AI systems is evolving rapidly. AI may help mathematicians explore more possibilities and automate some technical work, but determining the significance of a result will continue to require mathematical judgement.
The partnership does not guarantee that AI-generated discoveries will be correct or that every research community will adopt the same approach. Instead, it creates a framework for mathematicians to assess the technology, compare findings and develop common practices.
Ultimately, the initiative’s success will depend on whether it helps turn AI-generated output into mathematical knowledge that researchers can independently verify, understand, and build upon.