Summary
OpenAI says an independent mathematics advisory group will help review emerging AI-generated results and advise on research standards. The company says its new internal model has resolved more than 100 open problems, including the Navier–Stokes Millennium Prize problem.
OpenAI has announced that it is working with an independent mathematics advisory group to help guide how the company reviews and communicates mathematical results from AI systems. The group is hosted at the Institute for Advanced Study, and its initial members include nine mathematicians and researchers.
The announcement follows OpenAI’s account of rapid progress by a new internal model. The company says the model has resolved more than 100 long-standing open problems across most areas of mathematics, including the Navier–Stokes Millennium Prize problem, a landmark question concerning the mathematical foundations of fluid flow. OpenAI says it began training the model on August 28. The announcement does not name the problems or provide proof materials for those reported results.
OpenAI also points to concerns raised in a recent open letter about the negative effects of using solutions to open problems as a benchmark for AI systems. The company says these concerns prompted discussion about how to inform the mathematical community as the technology develops.
What the group will advise on
The group is expected to help OpenAI assess the significance of emerging results and advise on how to coordinate their dissemination. Its remit also includes academic and professional standards in mathematical research, and ways AI tools could support research and learning.
OpenAI says the group will operate independently. It will be able to offer advice without being asked, comment publicly on the company’s impact on mathematics, and change its membership. Members will not be paid by OpenAI. The group will not advise the company on how quickly to pursue its internal mathematical work.
The arrangement gives mathematicians a formal role in advising on how results are reviewed and shared, while leaving the pace of OpenAI’s internal work outside the group’s remit.
Initial members
The group’s initial members are:
- François Charles (ENS-PSL)
- Camillo De Lellis (IAS, GSSI)
- Timothy Gowers (Collège de France, Cambridge)
- Martin Hairer (EPFL, Imperial College London)
- Nikhil Srivastava (Berkeley)
- Ulrike Tillmann (Oxford, INI)
- Ravi Vakil (Stanford)
- Edward Witten (IAS)
- Melanie Matchett Wood (Harvard)
OpenAI describes the group as a first step in involving mathematicians in decisions about how AI capabilities in mathematics should be developed and used.