(FYI: There's a Ghost of Tsushima spoiler in here, if you care about that.)
It's pretty natural at this point to imagine some kind of diverging of the values and goals of those who do research math. The cheap first pass at a framing of this goes something like "mathematics vs AI". Indeed, probably the most combative people are those that speak out against the big bad AI companies, their lack of values beyond profit, their lack of care for the mathematics community. Perhaps they would not use the word "combative" - they are simply defending themselves and their sacred discipline from these new invaders and pillagers. I guess I use the word because it seems like they are ready for a fight.
I think the second pass is closer to correct, maybe something like truth vs understanding. At least, it's a lot harder to know which side is the right one by just looking at the words. The idea is relatively straightforward: by using AI, we can accelerate research in an extreme way; so fast in fact that we will forego understanding the proof of results so as to produce more and bigger results. The Truthers will replace understanding with verification, and with that, build much higher and much faster. Truths that we did not expect for decades will soon become a week-long project. Theorems, which were once the stars we sought, will become blips in new and grander constellations. We will really see, in a way never before, what really is. What is true. And because truth is what we're after, why would we stop?
The Understanders of course see it differently. The loss of understanding of the proofs and other building blocks of the theories we construct will ultimately lead to the end of mathematics. The AI companies offer faster research to get their foot in the door, but once their whole leg is in, AI will defeat conjecture after conjecture, slowly whittling away young math researchers until there is no next generation. And then only AI will know any mathematics, or at least, know any mathematics deeply. And as our systems become more and more complex in the use of and defense of AI, and more and more disciplines are displaced, who will be left that understands anything?
I see it like this. We used to build houses out of wood. We would find the trees, cut them down, slice them, prepare them. And we would build our homes and our monuments, by hand, as a group. We knew every part of the process, and it was our craft. Now we build with steel, but some kind of alien steel. We can tell how strong it is, but we don't know how it's made or where it comes from. But we build with it because building is what we do and the steel does what we want. So we use it, and teach it. But if the aliens start to leave or withhold, I'm not sure what we'll be left with.
Anyway, truth vs understanding, a fight over values. I credit Terry Tao for originally bringing values up as the essential next crisis for mathematicians, and from what I can tell he's an Understander. But not everyone is like him of course. In fact the cast of characters of the Schism is about as interesting as the thing itself. And I think once we sort it out a bit, we'll realize that even truth vs understanding is a bit off. Maybe code vs survival, or samurai vs ronin, or Napoleon vs Russia, or something else like that.
But anyway, we'll get to that. First the people in the game.
U1. On the Understanders side, the easiest to pick out are the people who care primarily about understanding. They are skeptical of AI, for the reasons I described above. But, if AI can be used in the plight of understanding, then it is not completely bad. We should be careful about it, but we should use it. One of their main questions is about how we are going to bring up the next generation. Before we gained understanding from studying, and proofs of open problems and theorems generally were a certificate of that understanding. The certificate is gone, and now it is crucial we figure out how to certify understanding in other ways.
U2. Other Understanders care more about the way mathematics is done, rather than some more pragmatic view about the best way to obtain understanding. This group is generally more explicitly anti-AI, and probably many of the Association for Human Mathematics people fit into this bucket, especially the ones in the AI-free caucus/inner circle thing. The main concern of this group is how to preserve the practice or discipline of how mathematics is done. Mathematics is a long-standing tradition, and to change it radically in favor of profit-seeking companies (regardless of if that is the intention or not) is to destroy it. Or at least is to destroy its soul, which to this group cannot be distinguished in any meaningful way.
U3. I think there's even a group in the Understander basket which is either artistic or hedonistic, depending on your view. They see math as an art, and do it for its own sake. Of course also to find out truth, but somehow truth is the thing they have to produce to continue being employed. The real goal of this group is the beauty of mathematics, within themselves. It is a solipsistic view, one that sees math as a joyful unfurling of some inner world. They speak to others about their work, hoping to transfer some of that joy that they experience. But it often cannot be transferred, and they go back to their office to experience it for themselves again. Their main question is harder to justify: will I still like doing math in the AI age?
T1. On the Truther side, you have some analogous types. The typical truther is similar to the typical understander, just with reversed value parameters. They see truth as more important than understanding, but not necessarily by so much. It's more like: truth is, at the end of the day, the goal. But leaving understanding behind will inevitably hollow out our ability to find truth. They are skeptical of the "pure AI mathematician"; an AI model spinning on its own in a data center somewhere, writing out more and more theorems towards some generalized human-less Knowledge. These people have seen incomprehensible truth enough times to know that it can't go too far without more substantiation. And because of that, they care about the next generation of mathematics much like the first group of Understanders do. But they are also seriously interested in how to integrate these new tools fast, and they see this question as inseparable from questions about the next generation.
T2. Beyond this, there is a Truther group similarly idealistic (though perhaps they would deny it) to the "practice of mathematics" Understanders. These people are a relatively small group, and maybe almost non-existent in the community as it is traditionally understood. They see math as producing truth, full stop. The understanding is secondary in a strong way, and it can be abandoned if it hinders progress towards the resolution of our most important conjectures and beyond. This group is small, but it will grow. First in the AI companies themselves, and arguably in some of the non-mathematicians that post solutions to famous problems. But soon perhaps math will be the last frontier of truth there is. A place where the power of AI meets the ever-increasing difficulty of bigger and bigger concrete questions, which may become rarer as AI gets better. "What is true?" There are many that are driven only by this question, and they may perhaps soon flock to math.
T3. And finally, there is an artistic truther as well. It is perhaps summed up best here, and differs from the artistic Understander in that it is less about finding and more about expressing. Their question - what is my truth? - is one that is somewhat common today in various areas across culture and society, at least in the US. What it shares with the corresponding Understander type is a certain sort of solipsism. The subjective truth of an individual is more important than whatever objective truth we could obtain from AI or from the power-wielding "greats" of the past. And their mantra: "the truth is already within you." Nothing hurt my faith in Yoda and the Force like the latest Star Wars trilogy, and I feel something similar about the vibes of this group.
Of course this doesn't exhaust everyone, but I would guess that most people fit roughly somewhere here. Or maybe straddling some, or in different ones depending on the month or the week. There are different values mathematicians hold beyond just truth or understanding, like maybe simplicity, or elegance, or communication, or clarity, or whatever else. But I think truth vs understanding is still the apparent banner of the values fight today, with many other discussions able to be at least partially subsumed.
So then, back to my cryptic claim that maybe even this banner is not quite getting at what we are fighting over, or maybe, what we are about to more seriously start fighting over. If there will be a true Schism in mathematics, it will run deeper.
Ghost of Tsushima is a story about the orphan samurai Jin in feudal Japan trying to aid in the fight against Mongol invaders. His uncle raised him to hold firmly to the honor ethic of the samurai code: look your enemy in the eye and respect them, in victory and in defeat, in life and in death. But as Jin fights, and does not see anything of this code in the invaders, he relies more and more on stealth and trickery. To stop the invaders, to survive - that is the goal above all else. And if this goal can be attained by losing some of that honor ethic and responding to the invaders in kind, then so be it.
In Ghost there are different personalities, different politics, different ideas about what to do; similar perhaps to the different types of mathematician described above. Families fight over these things yet stick together, and many would say that it's healthy. But the honor code is indisputable. To break it is not just to disagree, it is to abandon.
What I see most of all as the future breaking-apart of the mathematics community lies in that distinction. If it really is just truth vs understanding, then we can still live together. One group will push the frontier, and the other will come behind canonicalizing the new truths. Some people will have more disagreement than others with the other side, but there could likely be a general mutual respect which keeps the family together. And this is sort of how math departments were already before AI, honestly. It would be crazy to suggest that there was no rivalry whatsoever in terms of "what counts as real/proper math" between, I don't know, combinatorialists and number theorists, say. Communities are complicated unwieldy entities. As long as they share some overarching meta-goal or teleology, they can be held together.
And this is why code vs survival is much more likely to rip things apart. Because the teleologies are different. The code says that math is not math unless it is done right. And the survivor will fit round math into a square hole, so that it will change rather than die. These are incompatible because both see the vision of the other as death or heresy or choosing something else over the thing itself. To each group, the other is an alarm, a call to the realization that if this other group holds the reigns, then it will not be very long before math itself is destroyed. There cannot be mutual respect, because it is believed that each group is killing the thing that even enables the respect in the first place.
In Ghost of Tsushima, Jin and his uncle eventually fight to the death. There is no canonical ending in a way; once you "win" the fight, you can choose to kill him with honor, or to let him live, dishonored. Either way, you don't see him again after that.
What will break math apart is not the different types of Understanders and Truthers. What will do it is the argument over the holy discipline of mathematics. The U1 and T1 groups are not that different, nor are the U3 and T3 groups. But U2 vs T2 will be the start of the Schism, if it does indeed come to that. They are incompatible, and their fight will bear out one way or another, in hiring, funding, and whatever other friction points of this whole conversation. It will then force the other groups to pick a side, in interesting ways. If I had to guess, I'd probably say the 3's would generally go with U2 and the 1's generally with T2. But it'll be messier than that. And it'll probably even cut along mathematical sub-area lines honestly. The break will run deeper and more fractal-like than we can predict.
Anyway, I honestly hope some great Schism isn't what it comes down to. But for me, I am a survivor, and I will fight for math. Survival over honor: I would rather change it than let it stay "untainted" and be destroyed. Hopefully it doesn't come to that. But I think it may, and it is certainly already creeping in that direction. We'll see.