The Future of Mathematics and Mathematicians

A reader worried about the future.

I am writing this email as a young aspiring researcher/scientist. We live in a period of uncertainty and I have a lot of doubts about the decisions I should make. I've always been interested in mathematics and physics and I believe that a career in this area would be a fulfilling one for me. However, with the development of AI I'm starting to have some worries about my future. It is difficult to understand what is really happening. It feels like everyday these models are improving and sooner rather than later they will render us useless.

I know that the most important objective of the study of mathematics/science is understanding and that AI will not take that pleasure away from us. But still I feel that it will be removing something fundamental to the discipline. If we have a machine that is significantly better than us at solving problems doesn't science lose a part of its magic? The struggle that we face when trying to tackle the genuinely difficult questions is perhaps the most fascinating part of doing research.

I would very much like to read about your opinion on this subject. Will mathematicians/scientists/researchers still have a role to play in all of this? Will science still be an interesting subject to pursue?

There are two questions here, the future of mathematics and the future of mathematicians. Let's take them separately.

Looks like the same letter went to Martin Hairer and highlighted in a recent NYT article about the state of AI doing math and the too-early First Proof project. According to Hairer, "I believe that mathematics is actually quite ‘safe', I haven’t seen any plausible example of an L.L.M. coming up with a genuinely new idea and/or concept."

I don't disagree with Hairer but the state of the art can quickly change. A few months ago I would have said that AI had yet to prove a new theorem, no longer true.

It's too early to tell just how good AI will get in mathematics. Right now it is like an early career graduate student, good at literature search, formalizing proofs, writing paper drafts, exploring known concepts and simple mathematical discussions. But no matter how good it gets, mathematics is a field of infinite concepts, there will always be more to explore as Gödel showed. I do hope AI gets strong at finding new concepts and novel proofs of theorems, and see new math that might not have happened while I'm still here to see it.

For mathematicians the future looks more complicated. AI may never come up with new ideas and Hairer might be right. Or AI could become so good at theorem proving that if you give a conjecture to AI and it can't resolve it, you might as well not try. The true answer is likely in-between and we'll get there slower rather than faster. 

People go into mathematics for different reasons. Some find joy in seeing new and exciting ideas in math. Some like to create good questions and models to help us make sense of mathematical ideas. Some enjoy the struggle of proving new theorems, working against an unseen force that seems to spoil your proofs until you finally break through, and impressing their peers with their prowess. Some take pleasure in education, exciting others in the importance and excitement of mathematics. These will all evolve with advances in AI and the mathematician's role will evolve with it.

My advice: Embrace mathematics research but be ready to pivot as AI evolves. We could be at the precipice of an incredible time for mathematical advances. How can you not be there to see it? And if not, then math needs you even more.