OpenAI's announcement that its artificial intelligence agents had solved the Navier-Stokes problem, one of the most famous and difficult challenges in mathematics, has ignited what some mathematicians are calling an existential crisis in their field. The claim, made on 8 September, has also drawn accusations that the company misused human work and failed to give sufficient credit to researchers who were close to a solution.
The Navier-Stokes equations describe the motion of viscous fluids and are central to fields ranging from weather forecasting to aerodynamics. Solving the problem, which involves proving whether smooth solutions always exist in three dimensions, is one of the Clay Mathematics Institute's Millennium Prize Problems, carrying a $1 million award. Had a human mathematician cracked it, they would have collected both prize money and plaudits from their peers. Instead, the announcement has prompted a heated debate about the role of AI in mathematical discovery and the ethics of how these systems are developed.
The critiques levelled by mathematicians will be familiar to artists, office workers, or anyone else concerned about the impact of AI on their profession. OpenAI's results, like any arising from a large language model, depend on digesting work by human mathematicians. The OpenAI paper cites sources, but many believe the company did not give sufficient credit, especially to several mathematicians believed to be very close to a solution. Critics describe this as bad form that erases human ingenuity and hogs all the glory.
One of those mathematicians, Tristan Buckmaster, has raised concerns that work he was doing on Navier-Stokes using OpenAI's Codex model was seen by the OpenAI team. OpenAI has denied directly accessing this material, but could not rule out that data from Buckmaster's use of their products «helped improve our model». The admission has intensified worries about the boundaries between proprietary research and the data used to train commercial AI systems.
The episode has broader implications for the relationship between technology companies and the scientific community. Mathematicians argue that human insight remains vital to the field, not only for generating new ideas but for verifying and contextualizing results. AI models can process vast amounts of information and identify patterns, but they lack the ability to understand the deeper significance of a proof or to judge its elegance and importance. The fear is that tech firms, eager to claim breakthroughs, may overlook these distinctions and undermine the very human expertise that makes mathematics possible.
OpenAI has not released the full details of its claimed solution, and independent verification is still pending. The company's assertion that its agents solved the problem has been met with skepticism by many in the mathematical community, who note that the Navier-Stokes problem is not a computational puzzle but a fundamental question about the nature of mathematical truth. Without a rigorous proof that can be checked by human experts, the claim remains unsubstantiated.
The controversy also highlights the growing tension between the open exchange of ideas that has traditionally characterized mathematics and the proprietary, competitive nature of AI development. As tech companies pour resources into AI research, they are increasingly positioning themselves as leaders in scientific discovery. But mathematicians warn that this approach risks exploiting human labor without proper attribution and could discourage researchers from sharing their work openly.
For now, the debate over OpenAI's claim is far from settled. What is clear is that the incident has forced mathematicians to confront uncomfortable questions about their field's future and the role of AI within it. The outcome will likely influence how scientific communities engage with technology companies and how credit and recognition are assigned in an era of increasingly powerful machine learning systems.
