PRINCETON, NEW JERSEY / RankWire.AI / – OpenAI has initiated cooperation with a standalone group of mathematicians to review and disseminate results generated by AI in the field of mathematics. The Advisory Group on Mathematics and Artificial Intelligence is based at the Institute for Advanced Study in Princeton. Its role involves guiding emerging findings, setting academic standards, and overseeing the application of OpenAI tools in research and education. This group maintains independence from OpenAI, with its members not receiving any remuneration from the company.
The advisory panel can offer unsolicited advice and publish its recommendations publicly. It also has the capacity to comment on how OpenAI’s influence impacts the field of mathematics. Members oversee the composition of the group and may advise other AI organizations whose models have substantial effects on mathematics. Importantly, the group has no decision-making authority either at OpenAI or any other AI enterprise. Additionally, OpenAI has clarified that the group will not advise on the pace of its internal mathematical research.
Initially, nine mathematicians comprise the team: François Charles, Camillo De Lellis, Timothy Gowers, Martin Hairer, Nikhil Srivastava, Ulrike Tillmann, Ravi Vakil, Edward Witten, and Melanie Matchett Wood. Their affiliated institutions include the Institute for Advanced Study, Harvard, Stanford, Oxford, Berkeley, Cambridge, EPFL, and ENS-PSL. De Lellis and Witten serve as principal investigators for the Institute for Advanced Study project that supports this initiative.
OpenAI reports significant strides in mathematical research
OpenAI revealed the advisory setup while announcing new claims related to an internal artificial intelligence model. The company stated that training of this model began on August 28. Since then, OpenAI claims the system has solved over 100 longstanding open problems spanning various mathematical disciplines. The company characterizes these outcomes as work produced by its internal model. The advisory group notes that its immediate focus is on how OpenAI should coordinate the release of numerous significant results it has reported.
This announcement follows OpenAI’s release on September 8 of research addressing the Navier-Stokes existence and smoothness problem, one of the Millennium Prize Problems in mathematics. The company published an analytical proof along with a formalization using the Lean proof system. According to OpenAI, the proof was generated by its internal model, demonstrating that a smooth fluid can develop a singularity in finite time. The company also stated it does not intend to claim the Millennium Prize for this achievement.
External group emphasizes standards for review and dissemination
The formation of this independent advisory panel came after OpenAI approached some of its members about establishing an external review board. Instead of a formal board, the mathematicians created this autonomous group and extended invitations to additional experts. Their stated goal includes fostering responsible interaction between AI firms, mathematical research, and the broader mathematical community. The group aims to guide the proper presentation and dissemination of AI-generated mathematical findings. This effort aligns with ongoing discussions among mathematicians regarding how AI-produced research should be evaluated, credited, and shared.
Currently, the group is focused on assessing the mathematical results that OpenAI claims its internal system has generated. It is gathering input from mathematicians to develop recommendations on how these findings should be released. The group intends to publish its guidelines on its website for AI companies. Responses from mathematicians will remain confidential unless they choose to disclose them publicly. Final decisions regarding the dissemination of results will rest with the recipient companies, with the independent group serving as a conduit between mathematicians and AI industry stakeholders.