In 2026, there is consensus among the 38 public AAU universities that AI will provide a social benefit. That provision is often expressed as conditional, of course, on the important role of the university. But AI social benefit will arrive, and the university must invest in that future.

I could argue there is little evidence for this claim, but the real point is that “AI will provide a social benefit” is not a claim that can be defended empirically without deciding in advance what benefit means. So this is more of a normative assertion. It is also, importantly, a statement about the future. That makes it a story.

We can test the plausibility of the story on a subjective level. For example, we probably all agree that clean water is a social benefit. Do we agree that the internet is a social benefit? meh. Email? MS-Office? Smartphones? Automobiles? Coal factories? Maybe the question gets complicated in different ways. For example, it’s hard to construct the counter-factual on these older technologies.

The internet is the best analogy I can think of. It’s recent. It’s significant. It’s a computer technology (which makes it easier to compare to AI than the birth control pill for example). So is the internet socially beneficial? There is no easy answer to that question. Rather than coming to an answer to that question, we will just reach a point where the counterfactual of what the world would look like if the internet had never been invented is just speculative fiction. Perhaps we are already there and we still don’t know because the internet keeps changing.

If it’s nearly impossible to decide the social benefit of technologies in retrospect, then deciding their beneficence in advance makes little sense. A university’s motives for being “AI forward” is not mysterious. Universities are not acting independently in this regard. That is not to say their advocates are misleading people. The point is that these universities assert a broadly shared position on AI, so there’s no evidence to support the claim they could assert something else at this time. Sites within the university (departments, centers, faculty) might assert alternate positions but they do not bleed through into central institutional argument.

The question for faculty and students then is how we respond to the realization that our university really has no ideological option than to be AI forward. Obviously that ideological obligation will benefit certain segments of the university. And benefit the faculty and students who share this belief. This kind of circular reasoning is not unique to AI. If we said that the future will be research written in crayon and devoted millions of dollars to faculty who produced crayon-media research, there is little doubt that the institution’s future will have become crayons.

I don’t mean to equate AI and crayons except in that we cannot predict either’s future (except by inventing it as the techno-futurist claim goes). AI could become as “socially beneficial” as the internet, and universities would still look foolish with their out of scale claims. Universities made significant investments in the internet as a technology. And curriculum developed around it from Computer Science to Art. We made digital literacy and information literacy general education outcomes. The university was transformed by the internet. But we never had rhetoric or investment like this. The internet never became the driving mission of the university in the way AI is now presented.

So why? For AI to become the transformative social benefit that justifies this collective academic fervor, it might transform or even eliminate academia entirely. There’s no reason to assert that social benefit also means university benefit. We can already see that it means benefit for some in the university rather than others. Indeed many higher education AI optimists place their optimism on this transformation (though notably not elimination). Understandably they also narrate themselves in particular as benefiting in this transformation–implicitly if not explicitly.

Milan Kundera writes, “to be completely modern is to be the ally of one’s gravediggers.” We should be clear that even if one day people come to say that AI is/was a social benefit, this says nothing about our lives. Victorian factory workers worked long hours in horrible conditions, regardless of the long term social value of industrialization. Meanwhile, people continue to work in dangerous industrial conditions. Then we can consider the externalities of pollution and climate change. AI is also an industrial technology. As with the internet, the question of the social benefit of industrialization remains unanswered.

The consensus of public AAU universities is that, given that we proceed with care, AI will be a social benefit. This consensus is organizing university investment. I think that’s going to be a hard argument to make in the next 15-20 years. At the very best, that outcome will be ambiguous. If universities are entitled to predictions, here’s mine: the range of futures in which AI becomes the unambiguous social benefit currently advertised and the university survives in anything like its present form strikes me as remarkably narrow. .