Fields Medalists Warn AI Problem-Solving Race Threatens Mathematics
The Fields Medal, awarded every four years by the International Mathematical Union, is among mathematics’ highest honors. The dispute is not over whether artificial intelligence can solve difficult problems, but whether using famous conjectures as benchmarks mistakes an output for genuine mathematical progress. Mathematicians argue that lasting advances require conceptual insight, careful exposition, attribution and scrutiny by the research community — a process that also trains scholars to formulate new questions.
On Sept. 11, 2026, Terence Tao, Peter Scholze and Maryna Viazovska were among 25 Fields medalists who published “A Severe Misalignment of AI in Mathematics” on the Math and AI website. The declaration said large language models had made dramatic gains in recent months, while companies were rushing to announce solutions without allowing enough time for rigorous write-ups, validation or recognition of earlier work. It named no company or model and urged mathematicians, AI developers and society to act quickly so the technology strengthens rather than erodes creative research.
All Coverage
2 original reportsThe Backstory
The history behind this eventTerence Tao Warns AI Is Depleting Math’s Best Unsolved Problems
Rapid advances in generative AI and automated theorem-proving systems are pushing machines deeper into work that once required years of effort by mathematicians. Terence Tao, one of the world’s best-known mathematicians, argues that answers are not the only scarce resource: researchers must also formulate new questions that are deep, useful and capable of opening fresh lines of inquiry. If AI solves such problems faster than scholars can create them, the research ecosystem could lose a key engine of innovation.
Tao has warned that AI is consuming mathematics’ most valuable unsolved problems faster than the academic community can replace them. He proposed marking selected questions as “analysis required,” signaling that a bare AI-generated answer would not be sufficient without reasoning, interpretation and insight. The approach is intended to preserve the value of mathematical understanding and method-building even as powerful tools shorten the time needed to reach a solution.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →