Where to have your say on math and AI
Declarations, open letters, and more
A short list of the high-profile declarations and letters on AI in mathematics, who can sign or endorse each one, and what you are agreeing to when you do.
Here are the high-profile declarations about AI in mathematics specifically, that I know of, and how you can endorse, support, or sign each of them. Not all of them are declarations, and in some of them you can’t sign alongside the original signatories, but you can still add your name or your opinion. The counts are as of 23 September 2026.
A Severe Misalignment of AI in Mathematics
The Fields medallists’ declaration, published on 11 September and now signed by 27 Fields medallists (Avila et al. 2026). Practically, as long as you can agree with its central sentence,
The goals of the AI companies and the goals of the mathematical community are severely misaligned.
then you can endorse it. You sign in with ORCID or confirm an academic email address, and your name goes on the list of endorsers, which is at 7,922. Strictly, though, you are endorsing the whole text. Gowers, for example, didn’t sign because of a different sentence, that “solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight” — see A flood from within the community. See my view on this in progressive disclosure.
Open Letter to Sir Paul Nurse, President of the Royal Society
An open letter signed by 42 mathematical Fellows and Foreign Members of the Royal Society, asking the Society to convey to government and the media that the pace of AI development is an emergency (Fellows and Foreign Members of the Royal Society 2026). “Mathematicians of all types are invited to support the letter” through this form.
Practically, as long as you can agree that the issues it mentions aren’t hype, you can support it. E.g. it cites former employees of the large AI companies “estimating the probability of human extinction to be as high as 10 per cent over the next decade”. The letter is not saying this is true; it is saying that “such statements must not be dismissed as ‘hype’”.
Notably, Terence Tao didn’t sign it, because it says “None of us have any significant involvement with AI companies.” In his words, “I have worked in various collaborations with the AI industry and so was not eligible to be a signatory”, but he supports the letter (Tao 2026).
Leiden Declaration on Artificial Intelligence and Mathematics
Published on 2 June and endorsed by the International Mathematical Union (Alper et al. 2026). For this one, anyone can sign it, with ORCID or with any email address, and 4,180 people have. But this one is very long and has a lot of details: recommendations for individual mathematicians, for mathematical organizations, and for policymakers and AI companies. So it seems to imply that if you sign it, you agree with all the details. The two above are different: only Fields medallists and Fellows of the Royal Society could sign them, and the rest of us endorse or support them, which reads more like agreeing with the main point.
Open Letter about the Mathathon
Asks for the Caltech Mathathon, which starts on 30 October, to be suspended (A coalition of current and former Caltech mathematicians 2026). I commented on it separately in Suspending the Mathathon won’t stop the flood. You can add your name through a form, but only if you are eligible:
Signatories must either be members of the Caltech community (including undergraduates, graduate students, faculty, researchers, and affiliates of the Jet Propulsion Laboratory) or members of the mathematical research community (at the PhD level or higher).
Advisory Group on Mathematics and Artificial Intelligence
Not really a declaration, but a place for an opinion. The group is advising OpenAI on how to release the large number of results OpenAI says its internal model has produced (more on that in What if the button came with the understanding?), and it asks for input from the mathematical community through this form. Responses “will not be made public without your approval.”
A few more
Here are a few more high-profile ones you should know about.
Give Authors Control is a petition to arXiv to let authors decide if AI bots can read their papers: full access, or metadata only. arXiv is broader than math, but I consider it related to math. Anyone with an active arXiv account can sign it, verified by an institutional email address.1
The Association for Human Mathematics is not a declaration but an association. Membership is open to “all people in the mathematical research community at the graduate level and higher”, and it comes with commitments: not working with commercial AI companies, and not publishing mathematical texts generated by AI models. There is also an optional AI-free caucus, whose members do not use AI models in research.
References
Footnotes
The petition itself is kind of interesting. It reminds me of the copyleft movement: which one is more free? Now it is: which one is more open?↩︎