Earlier this year, 23-year-old Liam Price — who does not hold a degree in mathematics — drew attention by apparently resolving a mathematical problem that had been open for about 60 years after asking ChatGPT about it. In May, OpenAI announced that its artificial intelligence had refuted an 80-year-old conjecture also associated with Paul Erdős. These high-profile items attracted wide publicity and sparked debate within the mathematical community.
In response, 150 mathematicians recently released a joint 11-page statement urging governments not to accept the hype that AI systems can reliably solve complex mathematical problems; the report of the statement was covered by Futurism.
What the statement says and who spoke up
In the document, Ulrike Tillmann, Vice President of the International Mathematical Union, argued that the deployment of AI raises ethical and professional issues that cannot be ignored. Tillmann emphasized that the future of mathematical research should be guided by human judgment, fair and transparent practices, and the shared values of the global mathematical community.
Tillmann warned that commercial interests can distort how a model's capabilities are presented, and she encouraged policymakers to consult mathematicians before making AI-related decisions rather than relying on eye-catching headlines.
Leslie Ann Goldberg of the University of Oxford cautioned that current systems often produce plausible but difficult-to-verify answers that can be hard to distinguish from correct mathematical proof procedures. Because research builds on previous results, this is especially critical: the scientific community must be confident that earlier findings are truly correct.
Why the debate matters
The mathematicians' concerns are multifaceted: overstating technological achievements can mislead the public and policymakers, and model outputs that are not verifiable are risky if they become the basis for subsequent research. The statement also notes that the rise of AI is occurring at a time when funding for academic teaching and research is often shrinking, which could incentivize institutions to prefer automated tools over human expertise for cost reasons.
The authors further argue that regulation of AI should extend beyond mathematics to other domains with serious societal implications, such as military use and surveillance technologies.
Conclusion
The 150 mathematicians' joint statement calls for critical scrutiny of ambitious claims about AI and recommends that decision-makers consult domain experts. While AI can produce striking results, the reliability and verifiability of current systems remain central concerns, particularly regarding mathematical proofs and the research that depends on them.
