The AI-mathematics wave unsettles the profession as results outpace peer review
A steady drumbeat of AI-assisted results is forcing the mathematics profession to confront hard questions: how to referee proofs produced faster than humans can check them, how to credit machine contributions, and what a mathematician's role becomes. The disruption is institutional, not just technical.
The bottleneck has moved to human review. When a lab can ship ten formalised results in a manuscript and a model can disprove a conjecture in a day, the rate-limiting step is no longer producing mathematics but refereeing it — and the peer-review system was never built for a producer that works this fast. The institution, not the mathematics, is what strains first.
Credit and authorship are the next fault line. If a model finds the argument and a human formalises and verifies it, whose result is it, and how is it cited? The profession has no settled answer, and the answer it reaches will shape careers, hiring, and how the next generation of mathematicians is trained to work alongside machines.
The optimistic reading, echoing Tao, is that the role shifts rather than vanishes — toward posing the right questions, judging significance, and verifying, while machines handle more of the search. The unsettling part is that the transition is happening faster than the norms to govern it, which is why the profession's mood this year is a mix of amazement and anxiety.
36Kr — 2026 AI crisis for mathematicians: key impacts and challenges → · The Decoder — AI keeps cracking unsolved math problems, and mathematicians have mixed feelings →