The Formers — Summary

Margaret’s grandson Eli showed her an email at Sunday dinner. It was from his university: twelve paragraphs on academic integrity in the age of AI, prohibiting AI tools for any graded assignment. “They want me to write like it’s 1995,” he said. “Then I’ll graduate and every job will expect me to use AI constantly.” He was not wrong. The university was defending something. But it had not said what.

Margaret thought about why Eli was there at all. Not why the university admitted him. Why he went. She suspected he went for the same reason she went fifty years ago: to find out who he is. To be somewhere that takes the question seriously. To encounter people and ideas and difficulties that would make him someone different than he was when he arrived. The degree was a side effect. The going was the point. The university’s AI ban email was not a defense of that deal. It was a defense of the institution’s ability to continue grading essays.

Higher education has always been several things at once, held together not by design but by historical accident: knowledge transmission, credentialing, coming of age, social sorting, and something harder to name — the development of judgment, the formation of a mind that can enter a discipline and become a practitioner of it. Nobody came for the lectures. They came for the rest of it. The lectures were how you got access to the rest, because everything was bundled and you could not unbundle it. The knowledge transmission was the gate that opened onto the formation, and because it was the gate, institutions built elaborate systems for defending the gate, and eventually forgot that the gate was not the destination.

AI pulls the bundle apart. Knowledge transmission? AI does this better than lectures ever did. Credentialing? Employers are increasingly skeptical of what the degree signals when the knowledge component of every field is freely available. Coming of age, social sorting, formation? These still require presence, peers, time. AI cannot provide them.

The university’s AI ban rests on an unstated reason: institutions know how to test content knowledge, but they do not know how to test judgment. They do not know how to assess whether someone can tell when AI output is wrong, whether a student has become a thinker or merely a completer of assignments. So they ban the thing that makes content knowledge less valuable, in order to keep testing what they know how to test. This is a betrayal of the social contract. The student arrived under an implicit agreement: you come here to become someone, and we will provide the conditions in which becoming is possible. The institution is now defending its ability to assess whether students can produce unaided essays, which was never what anyone came for.

The honest policy would say something harder: AI handles content, so we are now explicitly in the business of developing judgment, and here is how we will do that. No university has written that policy. Most have not figured out what it would mean. Figuring it out would require admitting what the institution is actually for.

The distinction is between information and formation. Information is content that can be written down, stored, transmitted, retrieved. AI handles it. Formation is becoming a certain kind of person. Developing judgment. Learning not just what the field knows but how the field thinks. A person who can recite the history of economic thought but cannot ask an economic question, cannot recognize when an economic argument is inadequate, cannot judge when AI’s economic reasoning is wrong, does not know economics. They have been informed about it. You can be informed about philosophy. You cannot be formed as a philosopher by information alone.

What formers actually do: enculturation into a discipline is the first and least visible thing. The student learning history is not only accumulating historical knowledge. They are learning how historians think, what questions historians ask, what counts as evidence. This cannot be transmitted. It is absorbed through proximity to historians who are doing history. Judgment modeling is different from instruction — the professor working through a problem in real time, including the wrong turn corrected, the admission “I’m not sure about this.” AI can provide correct answers. It cannot model the process of reaching judgment. Standards enforcement is something personal: “This isn’t good enough for you,” not good enough in general, but good enough for this specific student who the professor knows can do better. Intellectual companionship is the office hour that changes a life — the professor who takes a half-formed idea seriously, bets on a future they may never see. AI cannot do this because AI has no stake in the student’s becoming.

What if forming and researching were actually different jobs? Researchers research. Formers form. Different hiring criteria. Different training. Different measures of success. Current professor training is a PhD in a discipline — deep content knowledge, minimal training in formation. The institution hires based on research, then expects formation to happen as a side effect. Sometimes it does. Often it doesn’t.

The university that survives will be the one that names formation as its core work, makes the contract explicit, and builds around keeping it. Not the one that bans AI to protect its assessment practices. The former is defined not by what they know but by what they develop in others: judgment, enculturation, the ability to enter a discipline and become a practitioner of it.

Eli will graduate. He will use AI constantly in his work. The question is whether the four years produced judgment or just credentials. Whether someone formed him or merely informed him. Whether he eventually tells someone he wants to study music, and whether anyone along the way created the conditions in which that admission was possible. The student who can name what they want, who has been given the space to find it, who has become someone in the four years — that student received what the institution promised. The student who leaves with a transcript but not a self has been informed but not formed. Most universities are producing the second kind. The contract says they owe the first.