We should remember to also look at what AI can give us, instead of just bemoaning humanity having to share another pedestal. Some examples: A free, live automatically and instantly updated database of every theorem ever proven and every proof of every theorem ever proven, with excellent explanations of every proof and great on-demand overviews of connections between all proofs and all areas of mathematics. (similar to a Chess opening explorer) No one will ever have to read a badly written math paper again. If someone manages to publish a badly written mat paper, you can just ask the AI to rewrite it better. AI will unslop the historical math literature. The AI can give you the historical context of all math you ever see. Instant and personalized excellent pedagogical journeys to and through any area of math that you want. Instant, excellent and personalized exercises, lectures and written material for every child and every adult that wants to learn any part of mathematics. For anything from Pythagoras' theorem to Andrew Wiles' proof of Fermat's Last Theorem. (similar to an AI Chess tutor, which are not yet well developed, but will be) Instant auto-formalization that means humans will never again believe that something is proven that isn't proven. Automatic applied math, expanding the capabilities of all other fields that ever use any mathematics. The biologist who doesn't know statistics can now be certain that they got their statistics right. This is more important than the rest, and pretty obvious, but somehow mostly forgotten in the discussion. AI labs can do some of this, but real mathematicians and regular people will have to participate in large numbers to really get this off the ground and this should be a public effort, not only a proprietary one. Like Wikipedia is. The AI won't know for sure if it is making good explanations without having a human there to tell it. And for some time, perhaps a long time, it will continue to be true that human mathematicians can help the AI make progress also on hard math topics like proving theorems. The role of mathematicians, operationally, has been to improve the literature by publishing papers. AI doesn't kill that, but it changes it. The purpose of mathematicians, operationally, is now instead to improve the AI. This is not terribly different from the previous purpose. You publish useful artifacts that serve the purpose of mathematics: to improve the literature, and, now, to improve the AI that we all use. No one has any use for luddite rebellions, but the good news is that there *is* much use for software engineers, regular people and, yes indeed, *mathematicians*, to participate in creating tools like this for the benefit of us all. A more useful and a broader mathematics for us all. That's what AI means to humanity. It's meaning is not captured by the noise that you'll find on X right now.