PRINCETON, NEW JERSEY / RankWire.AI / – OpenAI has initiated a partnership with a separate group of mathematicians to review and disseminate AI-generated mathematical discoveries. Hosted at the Institute for Advanced Study in Princeton, the Advisory Group on Mathematics and Artificial Intelligence will guide the evaluation of new findings, uphold academic standards, and oversee the utilization of OpenAI tools in research and education. The group functions independently from OpenAI, and its members do not receive compensation from the company.
This panel can offer advice unrequested by OpenAI, publish its recommendations, and publicly comment on the company’s influence on mathematics. Its members determine group membership and may advise other AI firms whose models substantially impact the field. The group holds no decision-making authority at OpenAI or any other AI organization. Furthermore, OpenAI has stated the group will not advise on the pace of its internal mathematics research.
The initial team comprises nine mathematicians: François Charles, Camillo De Lellis, Timothy Gowers, Martin Hairer, and Nikhil Srivastava. Ulrike Tillmann, Ravi Vakil, Edward Witten, and Melanie Matchett Wood complete the roster, representing institutions such as the Institute for Advanced Study, Harvard, Stanford, Oxford, Berkeley, Cambridge, EPFL, and ENS-PSL. De Lellis and Witten serve as principal investigators for the Institute for Advanced Study project that hosts this initiative.
OpenAI reports rapid advances in mathematical research
OpenAI revealed the advisory structure alongside new claims about an internal artificial intelligence model, stating that training started on August 28. Since then, the system has purportedly solved over 100 long-standing open problems across diverse mathematical domains. The company describes these achievements as outputs generated by its internal model. The advisory group notes its immediate focus is on how OpenAI should coordinate the publication of these substantial results.
This announcement follows OpenAI’s September 8 release concerning the Navier-Stokes existence and smoothness problem, one of the Millennium Prize Problems. OpenAI shared an analytical proof and a formalization via the Lean proof system, claiming its internal model produced the proof and demonstrated that a smooth fluid could develop a singularity in finite time. The company clarified it does not intend to claim the Millennium Prize for this work.
Focus on standards for review and publication
The advisory group was formed after OpenAI approached some of its members about establishing an external advisory board. Instead, the mathematicians created an independent panel and extended invitations for additional members to join. Their stated goal includes fostering interactions between AI companies, mathematical research, and the broader mathematics community. The group aims to promote responsible presentation and dissemination of mathematical findings, amid ongoing discussions among mathematicians about how AI-generated research should be evaluated, attributed, and communicated.
Currently, the group is concentrating on the mathematical results claimed to be produced by OpenAI’s internal model. It seeks input from mathematicians as it formulates recommendations on the appropriate release of these findings. The group plans to publish its guidelines on its website, with responses from mathematicians kept confidential unless approved for public release. Final decision-making authority lies with the companies receiving the advice, while the independent panel acts as an intermediary between mathematicians and the AI industry.