AI in mathematics gets $15M push from Simons, OpenAI and Anthropic
AI in mathematics is gaining fresh support as the Simons Foundation, OpenAI and Anthropic commit $15M in combined funding and resources to help researchers use and assess new tools responsibly.
AI may help mathematicians explore problems that once took years to work through. But when a machine proposes a proof, how can researchers be sure the answer holds up?
The Simons Foundation is teaming up with OpenAI and Anthropic on a $15 million initiative to help mathematicians evaluate and use AI tools while preserving the rigour that underpins their field.
Announced on October 15, 2026, the programme will fund research community initiatives, workshops and access to AI tools and computing resources. Its aim is not simply to put more AI in mathematicians’ hands, but to help researchers decide how to test, trust, and use these systems.
Helping mathematicians set the rules
The Simons Foundation, OpenAI and Anthropic will contribute a combined $15 million in funding and resources, with the foundation overseeing allocations and inviting researchers to apply through an open call for proposals.
Support will go towards workshops and community-led efforts that help researchers evaluate AI-assisted results, learn to use emerging tools and exchange views on shared standards. OpenAI and Anthropic will also provide AI tools and computing resources to foundation-supported researchers.
This assistance will extend beyond mathematics to include basic science. The focus on community involvement is significant. Mathematical research relies on precise reasoning and verifiable proofs.
AI systems may help identify patterns, suggest solutions or formalise arguments, but researchers still need reliable ways to check that the results are correct.
Building on proof-checking tools
Simons Foundation President David Spergel said the initiative aims to bring the mathematical community together as AI reshapes research. He emphasised that mathematics is not just a collection of proofs and theorems, but a shared pursuit driven by human curiosity.
The foundation is building on its existing support for mathematics, machine learning for science and Lean, a proof assistant that helps researchers verify mathematical reasoning in a formal, step-by-step manner.
Tools such as Lean offer one way to check whether mathematical arguments meet strict formal requirements. The foundation has also convened workshops, lectures and collaborations to help researchers examine AI’s role in the discipline.
Keeping human judgement at the centre
As AI systems become more capable of assisting with mathematical work, researchers face questions that go beyond whether a model can produce an answer. They must also determine how results should be verified, reviewed and incorporated into established research practices.
The $15 million initiative gives mathematicians a role in developing those approaches, rather than leaving decisions solely to AI companies.
AI may open new avenues for mathematical discovery. Establishing how its contributions can be trusted will be just as important as expanding what the technology can do.