š° Full Story
A coalition of mathematicians published the Leiden Declaration on Artificial Intelligence and Mathematics on June 2, 2026, warning that AI companies are using published mathematical work without consent and threatening core research values.
Endorsed by the International Mathematical Union and carrying signatures from prominent figures including Fields Medalist Peter Scholze, the 11āpage declaration identifies five key risks: AI systems producing plausible but unreliable proofs; models trained on published work failing to attribute human contributions; commercial incentives reshaping research priorities; the bypassing of peer review in favour of press timelines; and loss of mathematical autonomy.
The document urges individuals to disclose AI use and retain responsibility for correctness, asks institutions to adopt licensing that prevents nonconsensual training on published material, and calls on policymakers to increase oversight and invest in public computational infrastructure.
The declaration has drawn hundreds of signatories and follows highāprofile claims by AI labs about mathematical breakthroughs, intensifying concerns about attribution, quality control and the effects on earlyācareer researchers.








š¬ Commentary