AI-produced Yang–Mills existence proof placed under a physicists' shepherding committee
As part of the same 26 October 2026 announcement, OpenAI said its models produced existence proofs for quantum Yang–Mills theory, dated 9 September, and for quantum chromodynamics including coupling to fermions, dated 18 September, with a Lean formalisation of the Yang–Mills existence result completed on 16 October. Unlike the mathematics and complexity results, the artifacts were not published. On 16 October David Gross was asked to take custody of the work, and a shepherding committee of Michael Douglas, Gross as chair, Arthur Jaffe, Igor Klebanov and Nikita Nekrasov issued a statement describing the proof as an important step towards putting quantum field theory on a rigorous mathematical footing. The claim concerns existence only; the mass gap half of the Clay Millennium problem is not claimed, and the prize remains unclaimed.
Why It Mattered
Constructive quantum field theory in four dimensions has been open since the 1960s, and a rigorous existence proof for Yang–Mills theory with fermions would be the most consequential result in mathematical physics attributed to a machine. What makes this a distinct entry from the publicly released proofs announced the same day is the handling. OpenAI did not publish the manuscripts or the Lean formalisation, transferring them instead to a committee of senior physicists who decide when and how the work is released. That is a new institutional form: a private laboratory producing a result in fundamental physics and entrusting disclosure to an ad hoc body of named custodians rather than to a journal, a preprint server or the public record. The arrangement buys orderly expert review at the cost of verifiability, since no one outside the committee can examine the proof. For a history of how scientific norms adapted to machine-generated results, the contrast with the simultaneous GitHub release of the complexity and harmonic analysis proofs is the instructive part: the same lab, on the same day, used two incompatible disclosure regimes chosen in consultation with the relevant expert communities. The entry should be read as documenting a claim under embargoed review, not a confirmed theorem. Its eventual standing — confirmed, qualified or withdrawn — will determine whether October 2026 is remembered as the month machine proof entered fundamental physics or as a case study in premature announcement.
Who Built It
OpenAI; shepherding committee chaired by David Gross
Applications
- Theoretical Physics
- Mathematical Physics
- Automated Theorem Proving