OpenAI Establishes IAS Math Advisory Panel After Model Clears 100 Open Problems

OpenAI has announced the formation of an independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study in Princeton, New Jersey. The initiative follows the company's disclosure that an internal model trained since August 28 has resolved more than 100 long-standing mathematical open problems across various fields, alongside a solution to the Navier-Stokes Millennium Prize problem.

Anthropic engaged external mathematicians for pre-publication validation and produced verifiable Lean proofs rather than establishing a standing advisory board. The company provides direct model access to academic researchers and governs frontier capabilities via its Responsible Scaling Policy.

The rapid surge in automated mathematical proofs has triggered intense friction within academic circles. Earlier in September, an open letter signed by 25 Fields Medal-winning mathematicians warned that commercial artificial intelligence laboratories are threatening traditional intellectual research by rushing to solve famous mathematical problems for competitive benchmark positioning.

DeepMind's FunSearch paired large language models with automated evaluators to discover new constructions for the cap set problem in extremal combinatorics. AlphaProof and AlphaGeometry 2 achieved an equivalent score on the International Mathematical Olympiad using Lean formalization.

The newly established group consists of nine prominent mathematicians, including François Charles, Camillo De Lellis, Timothy Gowers, Martin Hairer, Nikhil Srivastava, Ulrike Tillmann, Ravi Vakil, Edward Witten, and Melanie Matchett Wood. Among the initial cohort, only De Lellis, affiliated with the Institute for Advanced Study and GSSI, was a signatory to the Fields Medalists' open letter.

Operating under a framework designed to ensure institutional distance from corporate directives, members will serve without financial compensation. The group holds autonomy over its internal membership roster, possesses the authority to publish unsolicited evaluations, and can independently voice public positions regarding the company's academic impact.

Despite its advisory mandate to assess the significance of emerging findings, assist in dissemination standards, and advise on academic integration, the body holds strict institutional boundaries. OpenAI explicitly stated that the group will not hold decision-making power over the company's operations and will have no authority to direct or decelerate internal research velocity.

The Institute for Advanced Study underscored this division in its own public statement, clarifying that while its affiliated experts will provide guidance, ultimate responsibility for all commercial and technical decisions remains strictly with OpenAI.

The advisory group is expected to begin reviewing emerging computational outputs and formulating communication standards for future mathematical discoveries. OpenAI stated that ongoing discussions remain focused on how advanced computational tools can be safely deployed to assist academic researchers without destabilizing established professional standards.