AI math breakthroughs draw pushback from researchers

OpenAI, Anthropic and other AI labs have announced breakthroughs over the past year on multiple long-standing mathematical problems, with reported progress that exceeded researchers’ expectations for current systems. The advances include solving one of the famed Millennium Prize Problems, positioning AI systems as increasingly influential tools in high-level mathematics.

The rapid pace has stirred backlash among mathematicians, who criticize a Silicon Valley approach described as “moving fast and breaking things” and worry that labs are pushing into the discipline without enough regard for mathematical norms. OpenAI has said it hopes to consult top mathematicians and learn from past mistakes, but skepticism remains over whether those commitments will shape future work.

OpenAI’s milestones have left parts of academia unsettled, with disputes focused not only on technical claims but also on attribution and process. Some mathematicians want OpenAI to prove it did not use their research, while critics argue that achievements expected to be celebrated have instead become flashpoints for trust between AI labs and the math community.