OpenAI announced the formation of the Advisory Group on Mathematics and Artificial Intelligence (AGMAI) in mid-September through a guest post on mathematician Terence Tao’s blog. The nine-member group, comprising elite mathematicians from top institutions, was presented as an independent body that would advise OpenAI and other frontier AI laboratories on reviewing and communicating mathematical research findings.
However, the announcement has drawn criticism from the mathematical community for creating widespread confusion about the group’s actual independence from OpenAI. While OpenAI repeatedly stressed that AGMAI operates independently and members are free to publicly challenge the company, many mathematicians initially believed it was an OpenAI-appointed advisory board. Martin Hairer, an AGMAI member at Imperial College London and EPFL, confirmed the group receives no financial or technical support from OpenAI and members face only standard confidentiality requirements. He acknowledged that OpenAI initially approached some eventual members about forming an advisory board, though the group ultimately decided to operate independently and invite additional participants.
The confusion surrounding AGMAI’s formation reflects broader tensions between OpenAI and the mathematical research community. The company previously faced significant backlash over how it announced a solution to a Millennium Prize problem, with disputes erupting over credit attribution and treatment of mathematicians. Hairer noted that OpenAI’s announcement wording, while technically accurate, did little to help establish the group’s independence, acknowledging the company’s skilled public relations approach.
The advisory group faces an immediate challenge: coordinating the release of more than 100 long-standing mathematical problems that OpenAI claims its unreleased internal model has resolved. This prospect is already generating concern among researchers about the impact of such a volume of breakthroughs on the field. Hairer acknowledged the group formed hastily, having begun operations only days before the announcement, with no established master plan in place yet. The endeavor underscores the tension between OpenAI’s eagerness to publicize its mathematical capabilities and the mathematical community’s skepticism about the company’s approach to collaboration and attribution.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI