Why Math Dumped OpenAI — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Fields medalists issued "A Severe Misalignment of AI in Mathematics" on September 11, 2026, signed by 25 recipients, publicly rebuking OpenAI's approach to AI in mathematics.
- The Association for Human Mathematics (AHM) followed on October 7, 2026 with its own "Statement on OpenAI's October 6 Release of Mathematical Documents," escalating the formal criticism of the company.
- Terence Tao, described as the greatest living mathematician, had been one of AI math's most visible proponents, adopting ChatGPT to learn Lean prover in October 2023.
- Lean FRO is credited with foundational, more honest contributions to mechanized mathematics, built with what the article claims is a fraction of 1% of OpenAI's resources.
- Kevin Buzzard noted in a 2019 talk that formalizing the proof of Fermat's Last Theorem would require an estimated $100 million and 10 years, and that no working mathematician would care because the proof is already accepted.
Why it matters: The two letters mark a formal break from OpenAI by a domain-expert constituency that previously championed AI-math collaboration — Terence Tao himself was an early adopter using ChatGPT to learn Lean. That mathematicians, once OpenAI's most natural allies, are now signing public rebukes signals that the company's top-down posture is alienating the very research communities its tools most directly depend on.
Ask SkimNews
