The Leiden Declaration on Artificial Intelligence and Mathematics, which has been published today, addresses the increasing role of artificial intelligence (AI) in mathematical research. It was developed following a 2025 workshop at the Lorentz Center in Leiden and in consultation with an international group of researchers.
The Declaration outlines a range of ways in which AI systems are being used in mathematics, including the formalisation of proofs. It highlights concerns about how AI may affect established practices in mathematical research.
The authors identify several potential issues. These include questions around the reliability of automatically generated results, the attribution of results that are produced using trained proprietary models, and the impact of new tools on publication practices and peer review. The document also discusses the role of commercial organisations in this area.
The Declaration sets out recommendations aimed at different groups. For individual researchers, these include disclosing the use of AI tools in their work, taking responsibility for ensuring results are correct, and ensuring appropriate citation of prior work. For professional bodies and funders, it suggests developing policies on AI use in publication and review, and maintaining existing standards of scrutiny. For policymakers, it raises issues relating to regulation, public investment in infrastructure, and the use of expert advice in assessing claims about AI systems.
Read the London Mathematical Society’s policy on AI and publications
Last updated: 2 June 2026