OpenAI formed an independent advisory group at the Institute for Advanced Study in Princeton to involve mathematicians in discussions of its AI research. The decision followed the company’s publication of results that drew a mixed reaction from the mathematical community.

The new group is called the Advisory Group on Mathematics and Artificial Intelligence. Its task is to serve as a bridge between OpenAI and the mathematical community, as well as to assess the significance of the company’s new results and help organize their publication.

The board was created shortly after OpenAI unexpectedly claimed to have solved the Navier–Stokes problem, one of the seven Millennium Prize Problems. Alongside the announcement of the new group, the company said that the same internal model had supposedly solved more than 100 additional open mathematical problems across most areas of mathematics.

Mathematicians demand more oversight

The rapid pace at which AI companies are publishing results on well-known mathematical problems has drawn criticism from researchers.

Earlier in September, 25 Fields Medal laureates signed an open letter warning that the race among AI labs for headline-grabbing mathematical results poses a threat to the intellectual work of mathematicians.

Against this backdrop, OpenAI decided to formalize its engagement with the scientific community. The company says the new group will give mathematicians an opportunity to participate in discussions about how AI results should be published and evaluated.

At the same time, the board will not become part of OpenAI itself. Its members will not receive compensation from the company, will be free to express their own public positions, and may offer recommendations to OpenAI on their own initiative. In addition, the group’s composition is controlled by its members themselves.

The board will not be able to slow OpenAI’s research

Despite its stated independence, the group’s powers are limited to an advisory role.

It will be able to assess the significance of the results obtained and take part in coordinating their publication, but it will not have the authority to influence the pace or direction of OpenAI’s internal mathematical research.

The company states directly that the board will not be responsible for making recommendations on how to “set the pace” of OpenAI’s own progress in mathematics.

The Institute for Advanced Study separately emphasized this limitation. The institute noted that the group may provide advice but does not have the authority to make decisions within AI companies, and responsibility for those decisions remains with the developers themselves.

Thus, the creation of the board does not mean that mathematicians have gained the ability to stop or slow the publication of OpenAI’s controversial results. Their role is primarily one of expert evaluation and consultation.

The board includes nine mathematicians

The group’s initial lineup includes nine prominent mathematicians. Notably, only one of them — Camillo De Lellis of the Institute for Advanced Study — had previously signed the open letter from the 25 Fields Medal laureates.

This means that the new board is not a direct continuation of the initiative by mathematicians who opposed the current race among AI companies to solve well-known problems.

Nevertheless, OpenAI now gains a permanent channel of communication with the professional mathematical community at precisely the moment when its claims about AI’s capabilities in mathematics are raising increasing questions about verification, significance, and the process for publishing such results.