OpenAI Withdraws 3 Publishes New Math Papers
openai
| Source: HN | Original article
OpenAI has withdrawn three mathematics papers, adding notices that detail gaps and link to archived versions, while revising fourteen related manuscripts.
OpenAI’s ambitious rollout of AI‑generated mathematics hit a quick correction on Tuesday, when the company removed three of the 722 manuscripts it had posted to GitHub just a day earlier. The withdrawn papers – including one claiming a proof of the rational Hodge conjecture for products of K3 surfaces – now display notices that explain the identified gaps and link to the archived versions.
The pull‑back follows a flurry of community backlash after OpenAI released the bulk of the work, describing each as a resolution or substantial advance on an open problem in mathematics or theoretical computer science. Mathematicians advising the firm have repeatedly warned that the dump does not constitute endorsement, and the Association for Human Mathematics has called for scholars to cease collaborating with OpenAI. As we reported on 8 October, the association labelled the release “power, not scholarship” and urged caution.
OpenAI’s response goes beyond the three withdrawals. The company announced revisions to 14 other manuscripts, tightening proofs, clarifying hypotheses and fixing an obsolete citation. The swift remedial action underscores the difficulty of vetting large‑scale AI output without traditional peer review, and raises questions about how research institutions will handle AI‑generated claims that may appear credible but lack rigorous validation.
What to watch next: OpenAI is expected to publish a detailed post‑mortem of the release, and the mathematics community is likely to push for clearer standards on AI‑authored submissions. Further scrutiny may also spill into adjacent fields, such as cryptography, where recent commentary has warned of AI‑accelerated advances. The episode will test whether AI can become a reliable partner in frontier research or remains a source of unvetted speculation.
Sources
Back to AIPULSEN