OpenAI has said that its AI agents solved the Navier-Stokes problem, one of the most demanding open questions in mathematics, in an announcement made on 8 September. The claim has not been verified, and mathematicians are still working out whether the result holds up.
The episode has drawn criticism that the company gave insufficient credit to human mathematicians. Tristan Buckmaster, a mathematician, has said that his earlier work using OpenAI's Codex tool may have been among the material the system drew on. OpenAI denies directly accessing that material, but it has not ruled out that such work influenced its models.
The dispute highlights a tension at the heart of the industry's pitch. Technology firms promote their models' mathematical ability, yet they continue to rely on human mathematicians to check the output. Confirming a novel tool or insight takes time and considerable intellectual labour, and that work is done by people, not machines.
Mathematicians as a group are broadly open to AI as a research aid, even as many have signed open letters objecting to the conduct of large technology companies. The debate over how humans should guide and interpret AI's mathematical work is ongoing.
The Guardian's editorial board invoked the view of the mathematician Bill Thurston, who argued in 2010 that mathematics produces clarity and understanding, not merely answers. Nestor Guillen has argued that the technology should be separated from the companies that build it and made more widely available. Until independent experts reach a verdict on OpenAI's claim, the question of who deserves credit for the work remains open.
We rewrote this story on 4 November 2026 after an editorial review.
Credit. Based on reporting for The Guardian. Is this your work? Write to us.
100
