The Leiden Declaration on AI and mathematics, endorsed by the International Mathematical Union (IMU), calls for action to address challenges posed by artificial intelligence in mathematical research. This community initiative aims to preserve mathematical values amid rapid AI developments affecting proof generation, attribution, and evaluation systems.

Current AI techniques produce unreliable mathematical arguments that are difficult to distinguish from correct proofs, threatening transparency and verification standards. AI models trained on published mathematical works often fail to properly cite human contributions, undermining traditional attribution systems. These developments put increasing pressure on the peer review process, jeopardizing standards for correctness and independent verification.

The declaration warns that automated mathematics threatens the autonomy of research, potentially prioritizing problems amenable to AI solutions over those of deeper significance. With increasing involvement of technology companies in mathematical research and university budgets under pressure, these trends risk permanently narrowing the scope and depth of mathematics as a discipline while affecting professional incentives and collaboration terms.