OpenAI plans to publish hundreds of previously unsolved mathematical problems on GitHub, a move that has drawn criticism from the academic community. Mathematicians argue that these findings should be presented as formal research papers with detailed methodology rather than social media posts or repository dumps, to allow for peer verification.
In August, OpenAI met with approximately 40 mathematicians to discuss the publication of AI-generated results. While participants recommended a traditional academic approach, researchers like Bryna Kra of Northwestern University claim these suggestions were ignored. OpenAI also reportedly failed to utilize community-standard repositories like Hexagon and Palomar.
Tensions escalated in September following an incident involving the Millennium Prize problem, which carries a $1 million reward. Mathematician Tristan Buckmaster accused OpenAI of misappropriating research he conducted with Levent Alpöge of Anthropic. Buckmaster alleged that an OpenAI employee, Sébastien Bubeck, exerted pressure regarding authorship.
The scientific community has expressed concerns regarding the concentration of computational power and what some describe as monopolistic behavior. While OpenAI has established an advisory group on mathematics and AI, researchers report no significant changes in the company's practices. Internal debates within OpenAI persist regarding the future of mathematics.
