OpenAI has convened an independent advisory group of nine prominent mathematicians to guide how the company reviews, communicates, and deploys emerging mathematical capabilities from its internal AI models. The group was announced on August 28, 2026, the same day OpenAI said a new internal model had resolved more than 100 long-standing open problems across most areas of mathematics, including the Navier–Stokes Millennium Prize problem.
The mathematicians — François Charles (ENS-PSL), Camillo De Lellis (IAS, GSSI), Timothy Gowers (Collège de France, Cambridge), Martin Hairer (EPFL, Imperial College London), Nikhil Srivastava (Berkeley, Simons Institute), Ulrike Tillmann (Oxford, INI), Ravi Vakil (Stanford), Edward Witten (IAS), and Melanie Matchett Wood (Harvard) — operate independently from OpenAI, receive no compensation from the company, and may publish their advice publicly. Their mandate covers assessing the significance of new results, coordinating dissemination, advising on academic standards, and shaping how AI tools can support mathematical research and learning.
Confirmed
- OpenAI has established an independent mathematics advisory group of nine named mathematicians.
- Members operate independently of OpenAI, are not paid by the company, may change the group's membership, and may make their advice public.
- Their mandate: assess the significance of new results, coordinate dissemination, advise on academic and professional standards, and shape how AI tools can support mathematical research and learning.
- The group will not advise on the pace of OpenAI's internal progress on mathematics.
- OpenAI says a new internal model has resolved more than 100 long-standing open problems across most areas of mathematics, including Navier–Stokes — a claim with no independent verification.
Unknown
- No independent replication or verification of the claimed 100+ solved problems or the Navier–Stokes resolution.
- No timeline or process for how the advisory group's recommendations will be implemented or made public.
- No detail on which other open problems were solved, beyond Navier–Stokes.
- No word on whether the internal model will be released, offered via API, or remain internal.
Our take
An unpaid, independent board of genuine mathematical stature is a real governance signal: the community gets a structural voice before deployment decisions are made. But OpenAI still controls the model, the pace, and the release gate — and the group explicitly cannot weigh in on internal progress speed. Its leverage rests entirely on whether OpenAI treats its counsel as binding or merely consultative. The test arrives the first time the group recommends against a release OpenAI wants to make.
