Fields Medalist Terence Tao Warns AI Is Undermining Open Science in Mathematics

On Sept. 10, 2026, Fields medalist Terence Tao published a lengthy social media post criticizing AI for undermining the centuries-old open science tradition in mathematics. He compared powerful AI tools to a steam bulldozer that flattens the difficulty terrain of mathematics, making it harder for mathematicians to judge which problems are worth pursuing.

The comments came after OpenAI announced late on Sept. 9 that its internal model had solved the Navier-Stokes equations, one of the Millennium Prize Problems. NYU mathematician Tristan Buckmaster publicly accused OpenAI of using massive computing power to jump ahead after learning of his team's promising approach, and of asking the team to remove an Anthropic collaborator from the paper. OpenAI researcher Sebastien Bubeck denied the allegation.

Tao said finding valuable questions is one of the most scarce resources in science, and that brute-force AI intervention takes away not only the answer but also the process of searching for it. He warned that current incentives are pushing researchers to stop sharing promising directions, which would reverse centuries of open science and cause long-term damage. He called on the academic community to establish new norms and reject black-box answers produced without genuine insight.

His view drew support from AI researchers including François Chollet and Gary Marcus. Marcus said AI companies should compete on producing new scientific insight rather than using compute advantages to claim solutions first.