OpenAI forms math advisory group after model breakthrough

OpenAI has created an independent advisory group on mathematics and AI at the Institute for Advanced Study in Princeton, following concern over the pace of progress by a new internal model.

Training for the model began on August 28, and the system reportedly solved Navier-Stokes, one of the Millennium Prize Problems, while also tackling more than 100 longstanding open problems across much of mathematics. The speed of the results prompted internal discussions at OpenAI over how to publish them safely and responsibly.

An open letter from 25 Fields Medalists warned against treating complex mathematical challenges as routine AI benchmarks. The scientists said competition among labs to announce high-profile results could undermine traditional intellectual work and academic standards.

The advisory group will evaluate the scientific value of new AI results, coordinate publication, advise on professional ethics, and weigh how AI tools should be integrated into education and research. Its members will serve without compensation and may publicly criticize AI’s impact on mathematics, but the group will not regulate OpenAI’s internal development pace or deployment decisions.