Mathematicians declare the end of an era – 27 prize laureates speak out against OpenAI

For years, leaders of the biggest tech companies have promised that AI will push out a huge chunk of the workforce. The scale of the disruption will supposedly vary from industry to industry, they say, but no field will be completely spared.

So far, no industry has seen mass displacement of workers on that scale, but some mathematicians are already convinced their field has entered a new era.

The trigger was OpenAI's announcement on September 8 that it had solved the Navier-Stokes existence and smoothness problem. The problem had stayed open for close to a century and sits on the Clay Mathematics Institute's list of seven Millennium Prize Problems. Instead of congratulations, the company got a wave of accusations from the scientific community.

The criticism boiled down to claims that OpenAI had essentially seized the work of two researchers who were preparing to publish a paper on a related problem. The company offered one of them, NYU mathematician Tristan Buckmaster, partial credit on the Navier-Stokes solution. At the same time, it tried to push his collaborator, Levent Alpöge, who works at Anthropic, out of the publication.

Last week, Andreas Thom of the Dresden University of Technology piled on with his own accusations. He noticed that OpenAI's solution to a different math problem, unrelated to Navier-Stokes, looked eerily similar to recent joint work he'd done with mathematician Gábor Kun. Thom asked OpenAI directly whether that work had made its way into the model, since he and Kun had been discussing the problem with ChatGPT.

OpenAI researcher Mark Sellke's response was short and blunt: "that did not happen." The company later posted a clarification on X, stating that neither the human researchers nor the AI agents involved had seen "any specific user data." It also added that it couldn't rule out that de-identified data derived from users' interactions with its products had helped improve its models.

That caveat is what made Thom start doubting the first answer.

When it really gets to [be] a more tense situation, they actually are much more careful in what they're saying.

Thom called Sellke's flat denial dishonest, saying the honest answer would have been to admit uncertainty. Buckmaster himself sounded far grimmer after the solution went public. In an interview with the Australian Broadcasting Corporation, he said the genie was out of the bottle and that the role of human mathematicians had changed for good.

Tristan Buckmaster, NYU mathematician, said:

I think it's pointless. Like, I think the game is up.

Thom agrees a line has been crossed, but sounds less resigned about it. He thinks the community will need to redefine the boundaries of its own field and rethink what math students are even taught. It's less about surrender, in his view, and more about rewriting the rules.

I think it will fundamentally change everything. We have to come together as a mathematical community to redefine our subject, to redefine what math students are supposed to learn.

Thom also raises a bigger question about credit. Models train on the work of countless people whose names never get attached to the output, leaving open the question of whether the resulting breakthrough should count as a single company's achievement.

He frames it as a question of whether results produced by AI should be counted as achievements of the entire mathematical community.

The whole creativity of the system is trained by human creativity.

Three days after the Navier-Stokes solution was published, twenty-seven Fields Medal winners – math's highest honor – signed a joint statement. It argues that the push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics, and to the mathematical community.

The signatories also noted that using AI to solve and publish math results raises severe attribution and plagiarism questions.

On Monday, OpenAI responded with a blog post announcing an advisory group formed together with independent mathematicians. The stated goal is to give the mathematical community a voice in how we move forward. How the group will actually operate in practice hasn't been spelled out yet.