The day after we showed our students Forevergreen, an Academy Award–nominated short film, something happened that I can’t shake.

The film did what good art is meant to do: it quieted the room, even a room of K–8 students. It made them lean forward, their attention held by a sudden, rare gravity. Through the seamless marriage of image, music, and plot, it bypassed social static to reach their minds and hearts, holding them, if only for a moment, in awe.

Later, in conversation with seventh- and eighth-grade boys, I told them that the piece had been created by more than two hundred artists over five years. Their response came quickly, not as flippant dismissal but as baffled practicality: Why? One student said what the others were thinking: “AI could have done it in five minutes.”

In those two small moments, the educational dilemma of our age revealed itself. Students have acquired more than a new tool. The technology is quietly remaking what they believe human work is, and what they believe a human being is for.

In the weeks since, I’ve heard a companion phrase everywhere. Show a new image, a new paragraph, or a new idea—something that once would have invited curiosity or debate—and someone will shrug: “That’s AI.” Not as a moral alarm, but as a conversational stop sign. A way to end the encounter before it begins.

Before going further, the tension deserves to be named.

AI is a threat to meaning. It is also a tool for redeeming time. Teachers use it to clear administrative undergrowth that crowds out actual teaching. Students use it to organize research, test an argument’s weak points, and move past the blank page into the harder work of thinking. These are not trivial gains. Time is finite. Attention is precious. If a machine can shoulder the scaffolding, more of both can go toward what matters.

The question, then, is not whether to use these tools. If we want to understand the role of generative AI in education, we must first understand the ends these tools serve. And that, in turn, forces a deeper, more uncomfortable question: What is education actually for? We must answer this not in the language of outputs and metrics, but in an older, more demanding sense: What kind of person is a school trying to form, and what practices form that person?

To answer that, it helps to return to something older than our panic: the nature of signs.

Augustine of Hippo makes a deceptively simple distinction. There are things, and there are signs. A thing is what it is. A sign points beyond itself. Words are signs. So are gestures. So are sacraments. In a classroom, student work functions the same way. An essay is not valuable as ink on a page; it is a sign that points to a student’s mind: his grasp of an argument, his habits of attention, his ability to name what is true and discard what is not. A lab report is not the point. It is a sign of whether a student can submit to reality, whether he can observe, revise, and be corrected by what is. The artifact is never the telos. It is evidence that something has happened in a person.

For a long time, that semiotic economy held because the sign was tethered to time. Work took time. Time left traces. Those traces could be read, imperfectly but meaningfully, by a teacher and a community. Even a mediocre essay carried a certain dignity. It bore the marks of limitation and struggle—the marks of a human being learning.

Generative AI breaks that tether. It manufactures signs without the slow formation those signs once implied. It delivers the look and tone of thinking without the ordeal of thinking. The change is more than the fact that the tool can produce an answer. It can produce a trace, and it can do so at scale. The very kind of object teachers have historically read as evidence of becoming is now cheap to generate.

Signs have always been detachable from their makers. A sentence can survive its author. It can be quoted, recopied, and recontextualized. The power of a sign is that it can travel, repeated in contexts the original speaker never intended. A signature can be reproduced. A text can be lifted out of its life. If a sign could not be repeated, if it were bound to one private moment of presence, it would not function as a sign at all. Meaning depends, in part, on this portability. Language works because it can outlive us.

So what, then, is new?

Not absence. Not detachment. Not even imitation.

What is new is the industrialization of detachment. Signs are produced at conversational speed, tailored to the user, and clothed in the social costume of personhood. The machine does not publish; it answers back. It adjusts. It apologizes. It seems to take responsibility while having no continuous self to bear the costs of its own words. It produces the kind of speech that, in ordinary human life, is bound up with vulnerability and obligation, yet it is structurally incapable of either.

Put differently, what is a permanent feature of signification—repeatability without the guarantee of presence—AI turns into an atmosphere. The classroom does not contain detached signs; it is saturated with them. And once detached signs become normal, students become unsure which, if any, sign can be trusted.

They may respond in two predictable ways: contempt and suspicion. One drains the work of dignity. The other drains the world of wonder.

This is not entirely new. Even Shakespeare saw what happens when the sign floats free of the substance. In Richard II, the deposed king stares into the nature of authority and discovers, too late, that he has confused the symbol with the thing:

For within the hollow crown
That rounds the mortal temples of a king
Keeps Death his court, and there the antic sits,
Scoffing his state and grinning at his pomp.

The crown is hollow not because it is fake, but because Richard wore the signs of sovereignty without the inner formation, character—even virtue—that would have made them durable. When the ceremony and pageantry are stripped away, nothing remains that can stand.

That is the risk we are running in education. The danger is not that students will use AI to cheat, though they will, but that we will gradually accept a world in which the signs of learning circulate without the formation that once made them trustworthy. Essays without writers. Voices without owners. Credentials without the slow ordeal of becoming.

The hollow crown is not the AI-generated paper. It is the graduate whose diploma no longer points to anything real.

For most of history, personal effort was the visible cost of meaning. A poem mattered because you could feel a person behind it: spontaneous insights, false starts, and long struggle, the refusal to let a line go until form and function united. The words were a kind of signature, more than ink: the trace of a life and a way of seeing. A painting mattered for the same reason. You could sense the time inside it, the hours and disciplines that leave their mark. Even when you could not name the technique, you could recognize the ordeal of seeing and making: an artist submitting to reality or rebelling against it, accepting limits or rejecting them, and then inviting others to see what he sees.

Schooling rests on the same premise. An essay is not valuable because it fills a page. It is valuable because it is a record of a mind learning to think, a heart wrestling with the other, a person who is real. The artifact is a sign. It gestures beyond itself.

But here is what we rarely say plainly: the crown being earned is not a credential. It is a self.

A student who has wrestled with a hard text, revised an argument under pressure, and failed and tried again is more than informed. He is more solid. He has learned, however imperfectly, to be answerable to reality rather than expressive of preference.

That solidity is the telos of education. Not information transfer. Not skill acquisition. Not even “critical thinking” in the thin sense the phrase has come to mean. The end of education, rightly understood, is the formation of a person who can stand somewhere: someone with a voice, convictions, and a relationship to truth that has been tested and is therefore owned.

Generative AI does not threaten that telos by being wrong. It threatens it by being easy. It offers the appearance of having thought without the ordeal of thinking. And the ordeal is not an obstacle to formation; it is the mechanism.

“AI could have done it in five minutes” is not finally about speed. It is a moral revaluation. It assumes that what matters is the output, not the ordeal; the image, not the seeing; the product, not the person becoming capable of making it.

Live in that assumption long enough and time itself begins to look like a mistake.

When time becomes suspect, education becomes vulnerable at its root. Schools have always insisted that the young endure what is not immediately gratifying: the resistance of a hard book, the humility of a wrong answer, the slow dignity of revision. They teach, often against every adolescent instinct, that truth is not produced by desire. It is discovered through attention, patience, question, critique, and resolution—a continuous, embodied encounter with what is real.

The AI environment trains the opposite instincts. It offers competence without apprenticeship. Fluency without understanding. It makes it possible to generate signs without the patience required to become the kind of person those signs once implied. A student who internalizes that pattern does not become lazier; he becomes less formed, less present, less able to bear the weight of difficulty without reaching for a prompt.

This is where C. S. Lewis offers an image that is difficult to shake. In The Great Divorce, the souls who arrive from hell are not monsters. They are shades—thin, translucent, barely there. They have spent so long avoiding the weight of reality, so long choosing the frictionless comfort of their own preferences over the hard country of what is, that they have become insubstantial. They cannot bear the grass of heaven; it is too solid for them.

The danger AI poses to students is not primarily moral in the narrow sense. It is ontological. A student who consistently outsources the struggle, who prompts rather than thinks, generates rather than wrestles, or accepts fluency as a substitute for understanding, is choosing the path of the shade. He is becoming less solid, less capable of conviction, less present to himself and to others.

And the tragedy is that he may not notice. The signs will still be there. The essays will be produced. The grades will be earned. The diploma will be conferred. The hollow crown will be placed on a head that was never formed to wear it.

This is why the question of when and how to use AI is not a policy question. It is a semiotic and pedagogical one. If the artifact no longer reliably signifies formation, schools must move closer to forms of learning that cannot be reduced to a generated trace.

That shift is practical. It looks like more in-class writing and problem-solving where the mind is visible in real time. It looks like drafts submitted with commentary on what changed and why. It looks like oral defenses in which a student must explain the choices behind his claims. It looks like lab notebooks that record failure, revision, and the stubborn confrontation with what is. It looks like teachers who assess not only a product but a trail.

The honest answer to when AI is useful for students is only after the productive struggle — when it serves their formation rather than replacing it. AI is useful when it clears away the mechanical after the student already understands the mechanics; when it helps a student see the gaps in her own argument rather than filling them for her; and when it accelerates the path to deep work because the student has already done the hard labor of finding her way.

But that answer only makes sense if we know what the hard work is for.

A school must be able to say, without embarrassment, that its purpose is not to produce graduates minimally trained for the economic and political order. Its purpose is to form people who can bear the weight of reality: people who can read with care, speak with honesty, revise under correction, and stand behind their words.

That is not a mystical claim. It is physical. Formation shows up in flesh and habit. It shows up in whether a student can sit still with a difficult paragraph without reaching for escape. It shows up in whether he can name what he believes and why. It shows up in whether he can endure the small humiliations of being wrong without collapsing into cynicism or defensiveness. It shows up in whether he can apologize without theatrics and repair without resentment. These are not “soft skills.” They are the contours of a person who can be trusted.

When a school succeeds in forming that kind of person, it produces something AI cannot replicate: a voice. Not a tone, not a brand, not a borrowed style. A voice in the deeper sense—the sense in which a person has something to say because he has become someone. His words carry weight because they point back to a self shaped by attention, constraint, failure, and love of what is true.

A student formed that way can use AI freely and wisely. He can delegate the scaffolding without surrendering the substance. He can move faster when speed is warranted, and slow down when slowness is the price of depth. He can use the tool without becoming its product.

Forevergreen gave my students a brief taste of that older gravity. Beauty felt costly, and therefore meaningful. The film reminded them, before they had language for it, that labor is not inefficient. It is dignified. The making of a thing is part of its meaning.

Then the room brought that gravity into contact with the new regime. Five years. Two hundred artists. Why?

Why read slowly? Why memorize? Why draft? Why become something when you can generate the sign of having become it?

The unsettling part wasn’t the question. It was how normal it felt.

That is the mark of a regime change: not when a technology appears, but when its values start to feel like common sense. When time feels like waste. When effort feels like error. When simulation feels sufficient.

Richard’s hollow crown is a warning, not a metaphor for despair. He lost his kingdom because he mistook the ceremony for the substance. We are in danger of making the same mistake, accepting the signs of education while quietly abandoning the formation that once gave them weight.

If schools forget that, they will still have essays. They will still have grades. They will still have endless, fluent signs. But those signs will no longer reliably point to a person. And once that link breaks, we will not only be unsure what student work means; we will begin to lose confidence in what words, promises, and judgments mean at all.

Chris Sibben

Chris Sibben serves as head of school at the Rivendell School in Arlington VA.