OpenAI forms math advisory group after AI solved 100 problems

OpenAI Forms Math Advisory Group After Problem Claims

OpenAI announced on September 21, 2026, an independent Advisory Group on Mathematics and Artificial Intelligence hosted at the Institute for Advanced Study in Princeton, New Jersey.

The company said its internal model has resolved more than 100 long-standing open problems across most modern areas of mathematics. OpenAI also claimed that this system solved the famous unsolved Navier-Stokes problem, a major Clay Mathematics Institute Millennium Prize challenge.

The announcement followed years of intensifying global competition among major artificial intelligence laboratories to demonstrate advanced mathematical reasoning as a public benchmark for new machine intelligence.

OpenAI said the group will advise on reviewing and communicating emerging mathematical results, helping assess their significance and coordinate dissemination. The company described the new panel as a formal link between itself, the mathematical community, and the wider general public.

The Institute for Advanced Study said the independent group will offer formal advice but holds no binding decision-making power at any major artificial intelligence technology company.

Advisory Group Independence and Initial Membership

The group will work unpaid, offer advice OpenAI did not request, publish its views freely, and control its own membership. Official organizers said this structure gives the panel a measure of independence from the company while keeping its role strictly advisory.

OpenAI officials said the independent group will not be responsible for advising the company on how to pace its future internal research progress on pure mathematics.

Nine prominent mathematicians have been named as initial members, including Dr. Camillo De Lellis of the Institute for Advanced Study. Only Dr. De Lellis also initially signed the earlier public letter from mathematicians who had won the prestigious Fields Medal.

The initial members also notably include Dr. Martin Hairer, Dr. Timothy Gowers, Dr. Edward Witten, Dr. Ulrike Tillmann, Dr. Ravi Vakil, and Dr. Melanie Matchett Wood.

The Institute for Advanced Study said it will give advice but does not have decision-making power at any artificial intelligence company. It added that each individual company will ultimately bear full responsibility for its own decisions and their direct public outcomes.

Mathematicians Challenge Claims and Attribution

OpenAI said its internal multi-agent model recently generated the Navier-Stokes solution over roughly 88 hours using approximately 10,000 autonomous artificial intelligence agents working continuously in concert.

Within days, 771 mathematicians signed an open letter warning about what they called "slop mathematics" and inadequate formal peer review. OpenAI then withdrew its recent one million dollar sponsorship of the prestigious Caltech artificial intelligence math contest on September 10.

On September 11, 2026, prominent mathematicians including 25 Fields Medal winners publicly published a widely circulated open letter titled "A Severe Misalignment of AI in Mathematics".

The letter said rushed announcements often leave no time for proper detailed formal writeups, careful isolation of new methods, or adequate citation of important previous work.

Dr. Jacob Tsimerman, a 2026 Fields Medal winner, launched the Mathematical AI Safety Institute to study advanced mathematics for AI safety. The new institute will operate entirely independently and plans to hire between 10 and 30 mathematicians starting in January 2027.