The Advisory Group on Mathematics and Artificial Intelligence was established on September 21 to provide guidance to AI firms regarding mathematical research and academic engagement. The group operates independently from AI corporations and its members serve on a pro bono basis. The organization is hosted at the Institute for Advanced Study in Princeton.
The initiative began after OpenAI invited several mathematicians to join an internal advisory board. Instead of accepting individual roles, the scientists formed an independent collective to include a broader range of experts. The group clarified that it lacks the authority to make binding decisions for OpenAI or any other technology company.
The group's primary objective is to advise OpenAI on managing the release of mathematical findings generated by its internal models. The organization explicitly stated that it does not verify the accuracy, origin, or volume of these claims. Recommendations will be released publicly, and the group intends to solicit input from the wider mathematical community.
The organization aims to establish a framework for how AI companies handle mathematical results, focusing on verification, disclosure, and academic collaboration. The group is not currently tasked with auditing OpenAI's specific claims. The impact of these guidelines on industry practices will be assessed following the publication of the group's initial recommendations.
