Overview
- OpenAI announced in a September 8 preprint that an internal multi‑agent AI system produced a proposed solution to the Navier–Stokes existence and smoothness question and provided a Lean formalization to check the proof.
- Two mathematicians, Tristan Buckmaster and Levent Alpöge, say OpenAI may have learned from their prior ChatGPT interactions and unpublished work, a claim OpenAI denies by saying no user inputs after July 3 influenced the system.
- Twenty‑five Fields Medal winners have publicly warned that AI complicates how credit and plagiarism are assigned in mathematics, and many leading researchers say machine‑checked proofs do not replace human vetting and explanation.
- The Clay Mathematics Institute said it will consider the Millennium Prize only after peer‑reviewed publication and community scrutiny, and independent peer review and validation of OpenAI’s result are still ongoing.
- Beyond this case, mathematicians fear fast commercial AI will change how research is done by shifting who produces results, raising questions about data provenance, recognition for human contributors, and the culture of mathematical discovery.