Fields Medalist and University of Chicago professor Yau Shing-Tung recently addressed the role of AI in mathematical research. He emphasized that humans must remain the primary subjects in research, while AI serves strictly as a tool. Yau criticized AI companies for withholding the chain of thought behind their models, often citing anti-distillation concerns as a justification for non-transparency.
Yau proposed that the mathematical community should develop and maintain its own open-source models to reduce dependence on private AI firms. He argued that these models must adhere to strict academic standards, including full traceability and the disclosure of reasoning processes. This approach aims to ensure that mathematical research remains rigorous and verifiable.
On September 11, Yau joined 25 other Fields Medalists, including Terence Tao and Peter Scholze, in signing a statement titled A Severe Misalignment of AI in Mathematics. The signatories warned that the current trajectory of AI development in mathematics is harmful to the field. They highlighted concerns regarding plagiarism, lack of proper attribution, and the absence of human oversight in validating new mathematical ideas.
