Mathematicians convened a working group that produced the Leiden Declaration, warning that OpenAI and other tech firms threaten core research norms. The declaration, endorsed by the International Mathematical Union, highlights how AI can generate plausible yet flawed proofs, flood literature with unverified claims, and skew funding toward commercial priorities.

The warning follows OpenAI’s claim that an AI model disproved an 80‑year‑old geometry conjecture. Critics argue the company withheld details about prompts, training data, and compute, making the result difficult to verify. The declaration calls for transparent disclosure, rigorous peer review, and safeguards against AI‑driven bias in hiring and publication.

Researchers fear that unchecked AI use will erode mathematical rigor and disadvantage early‑career scholars. The document urges professional societies to set guidelines, protect authorship rights, and preserve the discipline’s independence from commercial incentives. Ultimately, the mathematicians demand a return to peer‑reviewed standards before AI breakthroughs shape the field.

The debate underscores a broader clash between data‑driven industry and the human‑centered ethos of pure research.