I am not as optimistic as some posting on this site, either about the long term effects of AI on Maths or its effects on the world as a whole. I hope I’m wrong. Here are my worries:
1) First, let me state two major activities that I really enjoy when I’m doing maths and which provide much of my motivation to do it. Neither should seem strange to the majority of human mathematicians, whether ‘professional’ or amateur but I want to restate them clearly in this context:
a) Finding a problem from which a creative project is born, creating a suitable framework for my approach, checking it, querying it, tussling with the problems and (hopefully) eventually arriving at something satisfying. That’s a great feeling we must all know. It’s a reward for our creation, hard work and solitary reflection. It may also show a new path forward as well, perhaps to other existing problems, or reveal questions that intrigue me…
b) Discussing my ideas with other (need I say “human”?) mathematicians, partly to show off my creations (yes!) and partly to share and partly to get their input (new viewpoints, questions context etc) during the process. Maybe even find a link to what they’re doing in their work. As well as verbal discussion, I might send an email or a preprint on a partial result or maybe an account of my ideas or possible spin-offs. I might also put one on ArXiv or try to publish it in a journal.
The problem: unfortunately, as things stand with AI at the present time, I can’t help feeling that indulging in (b) may compromise my ability to enjoy (a): if someone takes my ideas and feeds them into an LLM or if the same LLM is directed to scrape the internet for papers on a certain topic that can help it solve a similar problem someone has proposed to it, and if it comes across my ideas, in that case I risk being deprived of much of the satisfaction described in (a): not only the “glory” of solving (probably a reprehensible desire and quite possibly unachievable anyway) but also the pleasures of tussling and creating.
So, at present I am thinking that I need to stop doing (b), or do it much less and more guardedly, for the sake of (a). Some take a more “selfless” attitude in this respect (or seem to), but I can’t be alone in feeling that I need to prioritise (a) otherwise I will experience frustration and, ultimately, lose motivation for either. This is not a good situation for me and, if many people react like me, it won’t have a good effect on the subject either. So how do we tackle it? Can we?
2) Secondly, I don’t think there has been quite enough focus in the discussions here on setting the problems AI is causing for academic mathematics in the context of the larger and – let’s say it clearly – much more worrying problems that it threatens to cause the human world as a whole. I would like to point out that these two issues are far from being unconnected. Neither is the first simply a minor corollary of the second: Obviously there will be no human mathematics if there are no humans to do it but, less dramatically, in both areas, the main threat is a loss of human control over human activities, whether that control is given away, lost by accident or stolen for ulterior motives by certain actors, possibly (but not necessarily) silicon ones. In this sense the present anxiety of mathematicians for their subject can be seen as a useful illustration of the larger problems we may all be facing pretty soon.
Received 13 October 2026.
Add to the discussion