On September 22, OpenAI announced the establishment of an "Advisory Group on Mathematics and Artificial Intelligence." The advisory group is based at the Institute for Advanced Study in Princeton and consists of nine mathematicians who will operate independently from OpenAI.
The more significant news is that OpenAI disclosed in its announcement: a mysterious model being trained since August 28 has successfully solved over 100 long-standing open problems in multiple mathematical fields, including the controversial Navier-Stokes Millennium Prize Problem.
Mathematicians are "shocked" by the progress speed, how to inform the academic community leads to internal discussions
OpenAI stated that the model's progress in the field of mathematics has surprised the company's internal mathematicians, which has sparked internal discussions about the best way to inform the academic community.
OpenAI believes that mathematics belongs to the field of basic science, and new mathematical discoveries made by AI may have wide applications. Therefore, the development and deployment of such capabilities require a more cautious approach. The company is currently studying how to deploy AI capabilities related to mathematics more broadly.
Previously, mathematicians Tao Terence, Deng Yu, Peter Scholze, and others jointly published an open letter expressing concerns about the potential negative impacts of using solving mathematical open problems as a benchmark for new AI systems. OpenAI said that these opinions indicate that AI companies need to engage in more thorough communication with the mathematical community.
Nine members include three Fields Medal winners, do not receive payment from OpenAI
OpenAI clarified that the advisory group operates independently from OpenAI, has the right to provide suggestions that the company did not request, publicly comment on OpenAI's impact on mathematics, and can decide independently to adjust the composition of members. The members do not receive payment from OpenAI. However, the advisory group does not provide advice on the pace of the company's internal mathematical research.
The initial members include François Charles (École Normale Supérieure, ENS-PSL), Camillo De Lellis (Institute for Advanced Study, Gran Sasso Science Institute), Timothy Gowers (Collège de France, University of Cambridge), Martin Hairer (École Polytechnique Fédérale de Lausanne, Imperial College London), Nikhil Srivastava (Berkeley, Simons Institute), Ulrike Tillmann (University of Oxford, INI), Ravi Vakil (Stanford University), Edward Witten (IAS), and Melanie Matchett Wood (Harvard University). Among them, Gowers, Hairer, and Witten are all Fields Medal winners; only Camillo de Lellis is also a signatory of the aforementioned open letter.
The advisory group will provide recommendations on the review and dissemination of emerging results, help OpenAI assess the significance of the results, and offer opinions on the academic standards of mathematical research. It will also provide recommendations on how OpenAI's tools can support mathematical research and learning.