OpenAI's Math Breakthrough: AI Solves 100+ Problems, New Advisory Team Forms

OpenAI has announced the formation of a new independent advisory group, the Advisory Group on Mathematics and Artificial Intelligence, which will be hosted at the Institute for Advanced Study in Princeton, New Jersey. This group is intended to provide mathematicians with a greater voice and input into OpenAI's ongoing math-oriented research. The company's announcement emphasized that the group will act as a crucial link to the mathematical community and the broader public, ensuring mathematicians contribute to the direction of future developments.

This initiative comes after a period of intense activity and some controversy, notably following the abrupt publication of a solution to the Navier-Stokes Millennium Prize problem. OpenAI further claimed that the same internal model responsible for this solution has also resolved more than 100 additional open problems across various areas of mathematics. The rapid pace and presentation of these results have drawn criticism from many prominent mathematicians. Earlier in the month, 25 Fields Medal-winning mathematicians signed an open letter, expressing concerns that AI laboratories are increasingly threatening their intellectual work by competing to solve famous mathematical problems.

True to its name, the new OpenAI group will primarily serve in an advisory capacity. Its responsibilities will include assessing the significance of new mathematical results and coordinating their public release. The group has been granted a measure of independence from OpenAI: its members, though unpaid, can offer unsolicited advice, make their views public, and control their own membership. However, a significant limitation is that the group will not have the authority to slow down or redirect OpenAI's ongoing internal mathematical research. As stated in OpenAI's blog post, "the group will not be responsible for advising us on how to pace our internal progress on mathematics," a point reiterated by the Institute for Advanced Study, which emphasized that their advice does not equate to decision-making power at any AI company.

Nine prominent mathematicians have been named as initial members of the Advisory Group. Notably, only one of these individuals, Camillo De Lellis from the Institute for Advanced Study, was also among the signatories of the Fields Medalists' open letter, highlighting a specific composition for this new advisory body.