Leading mathematicians fear AI is making their field dumber, and warn the rest of us is next

Details

Twenty-five winners of the Fields Medal, the highest honor in mathematics, published a joint statement warning that the AI industry's approach to mathematics is 'severely misaligned' with what mathematicians actually do. LLMs have become capable enough to solve 'major outstanding problems in many fields of mathematics,' and AI companies are treating those solutions as benchmark wins rather than occasions for deeper understanding.\n\nMathematicians argue that famous unsolved problems serve as 'landmarks and lighthouses' — what matters is the thinking that emerges during years of work toward a solution, not the solution itself. AI can produce an answer without the surrounding conceptual development: no discussions, no simplifications, no attribution to prior work. The signatories warn this floods the field with answers that 'destroy fertile ground instead of breathing life into new ideas' and raises 'severe attribution and plagiarism questions.'\n\nThe statement also addresses a specific controversy: OpenAI allegedly learned of a partial solution to a Millennium Prize Problem and reportedly attempted to publish its own solution first for publicity, ahead of the researchers who developed the partial solution. Signatory Terence Tao had previously warned of an AI-driven 'foundational crisis' in mathematics.\n\nFor the broader AI field, the statement carries a wider message: the process of intellectual work — not just its output — has value. AI speed and volume can erode the culture and infrastructure that make scientific progress possible.