Who reviews AI-generated mathematics?

OpenAI announced on September 21 that an independent group of mathematicians will advise on reviewing and communicating emerging results, and on tools for research and learning. Its announcement says an internal model has resolved more than 100 longstanding problems. That remains a company claim, not a finding independently verified by this article.

The advisory group’s own statement sets a narrower, practical task: helping coordinate the release of results OpenAI reports. It says its members are unpaid, its recommendations will be public, and it can advise other AI companies. Companies retain decision-making responsibility.

For mathematicians, the distinction matters: advice about releasing proofs is not certification that those proofs are correct. The group creates a channel for scrutiny and discussion; its formation is not itself a mathematical breakthrough.

References

• Dated announcement
• Group’s remit and current task