OpenAI convenes independent mathematicians panel after research release backlash

OpenAI announced on Monday a new independent panel of mathematicians to advise it and other AI companies on how they interact with mathematical research, including how new results are presented and released.

The nine-member group, called the Advisory Group on Mathematics and Artificial Intelligence (AGMAI), follows a string of high-profile OpenAI mathematical results that became a reputational crisis. Researchers told The Verge the panel is a good first step, but many said they were left with basic questions about what it will actually do, how much influence it will have and whether OpenAI will listen to it.

A celebrity-heavy roster

Members are drawn from institutions including Stanford, Harvard, Oxford and Cambridge, with several Fields Medals and MacArthur “genius grants” among them. The group will be hosted by the Institute for Advanced Study (IAS) in Princeton, New Jersey, the former intellectual home of Albert Einstein, John von Neumann and J. Robert Oppenheimer.

OpenAI said the group will operate independently, with the freedom to offer advice the company has not requested and to publish it. Members will not be paid, and the group can change its own membership. “Its value depends on its members being able to exercise their own judgement and challenge ours,” the company said.

Questions about selection and influence

It remains unclear how much influence the group will have, or why it was formed. Simon Machado, a researcher at ETH Zurich, called the whole thing “shrouded in mystery” and questioned how the nine members were selected. The group’s website says it “came together after OpenAI approached some of its members about establishing an external advisory board”, but the mathematicians decided to launch an independent group instead.

Advisory group member Martin Hairer, a professor at Imperial College London and EPFL, published a blog post on Tuesday responding to what he called a “torrent of misinformation”. He stressed that not all members had been contacted by OpenAI, and that the group is “genuinely independent” of the frontier AI labs.

Francesco Fournier-Facio of Heriot-Watt University said greater engagement with mathematicians “can only make things better”, but that the panel “feels like an ivory tower”. Kevin Buzzard, also at Imperial College London, asked why such an elite body was needed: “It wasn’t proofs of hard theorems, it was better understanding of our subject.”

What comes next

Announcing the panel, OpenAI also said its unreleased model — the one that resolved the Navier-Stokes Millennium Prize problem — has now “resolved more than 100 long-standing open problems across most areas of mathematics”. Coordinating the release of those results is the group’s first priority, Hairer said. Members would “be very naïve to believe that the AI labs won’t try to spin whatever we say”, he added.