Yu Deng calls for transparent reasoning chains and open models in mathematics
In a keynote reported on October 26, Fields Medalist and University of Chicago mathematics professor Yu Deng said people should remain the principal agents in AI-assisted research and AI should serve as a tool. He argued that major AI-generated mathematical proofs should disclose their reasoning chains, criticized AI companies for citing anti-distillation concerns when withholding them, and proposed open mathematics models governed by academic norms covering data use, disclosure and traceability. Reports also said that Yu Deng, Terence Tao, Peter Scholze and 22 other Fields Medalists were among the 25 initial signatories of a September 11 statement titled “A Severe Misalignment of AI in Mathematics,” which warned of a gap between AI companies’ goals and those of the mathematics community.
Deng’s proposals directly address whether AI-generated mathematics should meet academic standards of scrutiny, attribution and traceability. The joint statement indicates that concerns about the relationship between AI companies and mathematicians are shared by a broader group of Fields Medalists.