The Association for Human Mathematics (AHM) has criticised OpenAI’s approach to mathematical research and urged researchers not to work for the artificial intelligence company. The statement followed OpenAI’s release of a large collection of mathematical results generated by one of its advanced internal AI models.
The release included 722 manuscripts covering 372 groups of results, according to the report. OpenAI said the work contained solutions to, or significant progress on, hundreds of open problems in mathematics and theoretical computer science.
AHM challenged the decision to publish such a large volume of AI-generated research, opening its statement with the remark, “Mathematicians did not ask for this work to be done.” The group also referred to allegations of plagiarism, copyright infringement and trademark dilution that have been made against OpenAI.
The dispute highlights a wider debate about how AI systems should contribute to scientific research, how mathematical discoveries should be evaluated and how credit should be assigned. AI models can help explore complex problems and generate potential solutions, but researchers still need to examine whether proposed results are correct, reproducible and genuinely novel.
Publishing a large number of manuscripts also raises questions about the burden placed on mathematicians who may be asked to verify or assess the results. Independent review is particularly important when claims involve longstanding open problems, where a convincing-looking argument may still contain errors or gaps.
AHM’s call for researchers to reject OpenAI reflects its objections to the company’s approach. The available report does not provide the group’s complete statement or OpenAI’s detailed response to the criticism.
The disagreement illustrates the tensions emerging as AI becomes more involved in research, including questions of scientific responsibility, transparency, attribution and the role of human expertise in validating machine-generated work.

AdvertisementThe Puranic