07-26-2026, 11:11 PM
Thoughts about the Leiden Declaration
BY Timothy Gowers
Summary
In this thoughtful response to the Leiden Declaration on Artificial Intelligence and Mathematics, Timothy Gowers explains why he chose not to sign the declaration despite supporting its overall aims. He praises the document for stimulating an essential discussion about AI's impact on mathematics but argues that it expresses several claims with greater certainty than he is comfortable with.
Gowers reflects on the traditional values of mathematics—proof, transparency, attribution, independent verification, and expert judgment—and agrees that they are worth preserving, while questioning whether all current practices, especially the evaluation of mathematical significance and merit, should remain unchanged. He suggests that AI may challenge long-held assumptions about authorship, understanding, and the nature of mathematical research, making it important to distinguish between core values that deserve protection and conventions that may need to evolve.
Rather than offering firm prescriptions, Gowers advocates humility, openness, and careful reflection as the mathematical community adapts to increasingly capable AI systems, arguing that the conversation should remain flexible as both the technology and its implications continue to develop.
ARTICLE
BY Timothy Gowers
Summary
In this thoughtful response to the Leiden Declaration on Artificial Intelligence and Mathematics, Timothy Gowers explains why he chose not to sign the declaration despite supporting its overall aims. He praises the document for stimulating an essential discussion about AI's impact on mathematics but argues that it expresses several claims with greater certainty than he is comfortable with.
Gowers reflects on the traditional values of mathematics—proof, transparency, attribution, independent verification, and expert judgment—and agrees that they are worth preserving, while questioning whether all current practices, especially the evaluation of mathematical significance and merit, should remain unchanged. He suggests that AI may challenge long-held assumptions about authorship, understanding, and the nature of mathematical research, making it important to distinguish between core values that deserve protection and conventions that may need to evolve.
Rather than offering firm prescriptions, Gowers advocates humility, openness, and careful reflection as the mathematical community adapts to increasingly capable AI systems, arguing that the conversation should remain flexible as both the technology and its implications continue to develop.
ARTICLE
┌────────────────────────────────┐
│ KONSTANTINOS MICHAILIDIS │
└────────────────────────────────┘
│ KONSTANTINOS MICHAILIDIS │
└────────────────────────────────┘