Students Wrote a Coherent National Policy on AI

Students Wrote the Rules. Schools Still Write the Script.

In late July, roughly a hundred high school students from across the country gathered at the Edward M. Kennedy Institute in Boston, formed a mock Student Senate, and produced a coherent national policy framework for artificial intelligence in K-12 education. They called it the STUDENTS FIRST Act. They debated, compromised, amended, and passed it by a wide margin. The resulting document is now being circulated to school leaders.

It is a serious piece of work. That fact deserves to be stated clearly before any critique.

The Value of What They Created

The Act is more thoughtful and more balanced than most adult-generated school AI policies. Students insisted on early and ongoing AI literacy covering bias, privacy, misinformation, environmental impact, and how the systems actually function. They prohibited the use of AI for the act of writing while allowing editing, brainstorming, and studying after eighth grade. They required citation of permitted use. They banned AI on graded assessments. They created a verification-of-mastery standard: if AI was used in drafting, students must be prepared to demonstrate understanding through discussion, handwritten work, or oral defense. They limited the authority of AI detectors and required human investigation. They prioritized human relationships over machine mediation in counseling contexts. They protected teacher judgment while still demanding transparency from teachers about their own AI use.

These are not trivial provisions. They reflect lived experience with the tools and a genuine effort to protect thinking rather than merely police outputs. In a landscape where adults have largely failed to produce coherent national guidance, the students succeeded in drafting something practical and usable. That achievement should be recognized.

The Limits of the Frame

And yet the document remains largely inside the traditional operating logic of schooling.

There is a reason schooling is so ubiquitous in the modern world. It matches deep human wiring. We are a social species that evolved in small groups where clear roles, hierarchy, belonging, and predictable behavior were survival advantages. Schooling scales those ancient patterns into mass societies. It produces controllable, structured outcomes: large numbers of young people sorted, supervised, and trained in the habits of compliance that large organizations have historically needed. The system is not primarily designed to cultivate independent judgment or self-direction. It is designed to manage groups and generate readable signals of trainability.

The STUDENTS FIRST Act works inside that reality. Its strongest energies go toward defining authorized versus unauthorized use, establishing detection and investigation procedures, clarifying disciplinary pathways, requiring published rules, and protecting the integrity of graded work. These are necessary short-term guardrails. They do not transfer genuine directorial power to students over their own learning. The school still sets the terms. The school still designs the literacy curriculum. The school still decides the boundaries of acceptable AI use. Students are expected to comply with a clearer and more carefully negotiated set of rules.

Full student agency would make schools much harder to manage. It would require adults who can tolerate messiness, support self-direction, and evaluate growth that does not always appear in standardized outputs. Most professional incentives and accountability structures are not designed for that work. Historically, the labor market has not demanded it either. The skills of independent judgment and free inquiry have often been treated as secondary to the more immediately rewarded capacity for reliable compliance.

That calculus is beginning to change. Generative AI is automating large categories of the very compliant work that traditional schooling trains people to perform. The scarce capacity is increasingly the ability to decide what is worth asking, to judge machine output, to notice when official stories diverge from what is actually happening, and to direct tools rather than merely operate inside them. Agency is becoming more instrumentally valuable precisely because compliance is becoming cheaper to automate. The lag between this shift and the still-dominant institutional signals creates the difficulty. Parents, students, and school systems can still look at current hiring practices and conclude that conformity remains the safer path. In many individual cases, for the next several years, they will not be wrong. The longer-term risk is a generation trained primarily for a form of work that is being automated away.

The Dramatic Irony

Here is the sharpest irony. A group of students demonstrated, under real deliberative pressure, the capacity for sophisticated collective reasoning about the systems that govern them. They produced a policy document of genuine quality. That document is now being handed to adult institutions that will almost certainly treat it as another external policy artifact to be reviewed, adapted, and implemented for students rather than with them.

Most districts that receive the text will run it through the familiar adult channels: administrative review, legal counsel, curriculum committees, and board processes. Student involvement, if it occurs, will typically be limited to after-the-fact feedback or symbolic representation. The same institutional logic that routinely excludes students from the design of their own learning plans will reassert itself. The creation of the Act proved what students can do when given a genuine deliberative space. The probable implementation path will demonstrate how rarely that space is maintained once the product returns to ordinary school structures.

This is not primarily a failure of goodwill. It is the default operating pattern of a system built to manage large groups in ways that feel natural to our social wiring. When student voice is invited, it is usually for input or legitimacy. Once the content exists, control of process reverts to those who hold formal authority and institutional accountability. Even a student-authored policy that tries to protect learning and relationships still travels through channels that treat students as the regulated population rather than as ongoing co-authors of the conditions under which they learn.

What Would It Mean to Take the Capacity Seriously?

The deeper opportunity is not simply better rules for the existing game. It is the recognition that the students who created this document already demonstrated a form of agency the system is structured not to fully use. Taking that capacity seriously would mean treating student participation in the ongoing design and revision of AI policy as a permanent feature rather than a one-time event. It would mean experimenting with genuine student authorship of individual learning plans. It would mean adults who understand their role as supporting the development of judgment rather than primarily managing compliance.

None of that is easy. It collides with how schools are currently organized and with the still-powerful short-term signals of the labor market. It also collides with the deeper reason schooling feels so natural: it fits the ancient human preference for structured belonging and controllable group outcomes. But the alternative is to celebrate a student-produced document while continuing to operate as if the most important educational work is teaching young people to follow clearer rules inside a system whose underlying bargain is eroding.

The students showed they can think carefully about the tools that are reshaping their education. The open question is whether the institutions that receive their work are prepared to treat that capacity as something more than a temporary input.

Why School Will Survive the Death of the Diploma (At Least for Now)

This continues the argument of "Why School Is the Same Everywhere, and the Revolution Never Comes," which ended with a claim I deferred: that AI will not revolutionize education, but may produce a crisis in what education is actually for. Here is that longer argument.

In the last post I tried to show why the education revolution never comes. The short version: revolutions are always promised to the story of school--individual flourishing, unlocked potential--and stories don't run schools. The operative functions do: sorting, credentialing, custody, the management of the young at scale. Film, radio, television, computers, and the internet each attacked the story, which is why each was so easily absorbed. The story flexed, the machine was domesticated, and the room remained the room.

But notice what that account implies. If revolutions fail because they aim at the wrong layer, then the institution has a right layer--a place where it can actually be hurt. It is simply a place no reformer ever aims, because reformers, by definition, are people who believe the story.

So let me make the distinction explicit, because it is the thesis of this post. In education's long history with technology, the promised revolutions have all been narrative-layer events: aimed at the story, announced loudly, and absorbed without residue. If real trouble ever came, it would have to be an operative-layer event--arriving not through the story but through one of the functions, and therefore not through any door the institution watches. For a century we have watched technologies fail at the first. I think AI is the second: the first technology in the sequence that does not attack the story of school at all. It strikes, almost by accident, the one load-bearing mechanism underneath, and nobody was even aiming at it.

The Severed Signal

I have made the full version of this argument elsewhere, so here it is in compressed form.

The bargain at the center of schooling has always been compliance for credentials: produce the assignments, sit the years, and receive a signal that employers trust. The signal worked--not because it measured learning, which it never did, but because the compliance it certified was expensive. Someone had to actually sit there, do the reading, grind out the essays, and show up for years. The cost of the performance was the information in the signal. It said: this person will persist, follow instructions, and finish what they start.

AI makes the compliant output free. And a signal that is free to produce carries no information. The diploma now certifies that a student had access to a chatbot, which is to say it certifies nothing. This is the operative-layer wound: not to instruction, which the school will happily automate, but to certification, which was the product.

Here is what should follow, logically: employers stop trusting the credential, students stop paying for it, and the institution built on selling it comes apart.

Here is what I think will actually happen: nothing. For quite a while. And the reasons why nothing happens are, I think, the most useful part of this whole analysis, because they tell us what to watch for.

Why Nothing Visibly Breaks

A credential has two kinds of value, and we habitually confuse them.

The first is informational: the credential tells you something true about its holder. That is the value AI just destroyed.

The second is coordinational: the credential is the thing everyone has agreed to use, and the agreement has value independent of the information. Employers screen by diploma because other employers screen by diploma, because HR software screens by diploma, because the whole hiring apparatus is built around it. Colleges require grades because the next stage requires them. No single actor can defect first--the employer who unilaterally stops requiring degrees has no replacement filter and looks reckless; the student who unilaterally skips the credential is gambling alone against the whole coordinated system (which has been the homeschooler's dilemma). So everyone keeps honoring a signal that everyone increasingly suspects is empty, because the honoring, not the signal, is what they actually need from each other.

This is why the crisis will be slow, and why its slowness will be mistaken for stability. We have run this experiment before, on a smaller scale: grade inflation. The informational content of grades has been decaying for decades--everyone in higher education knows it, employers know it, students know it--and the system has sailed on regardless, because grades kept their coordinational function long after they lost their informational one. The credential is following the same path, just faster and further.

So the first prediction is really a warning about expectations. We imagine a crisis announcing itself the way crises do in the movies: a visible, unsolvable problem, out in the open, demanding a response. This one will not look like that. Enrollment will continue. Diplomas will be conferred. The story will be told at every graduation. The absence of visible breakdown is not evidence of health--it is simply what this kind of failure looks like from the outside, because the coordination holds the surface together long after the substance is gone. A fiction of this size does not collapse because someone exposes it; I said in the last post that fictions never do. It persists, hollow, until the thing holding it up is quietly removed. Which means that if you want to see the crisis, you cannot wait for it to look like one. You have to know where to look.

The Visible Phase: The Enforcement Spiral

There is one place where the crisis is already visible, if you know what you are looking at.

Watch what the institution spends money on. Proctoring software. AI-detection tools that don't work, replaced by more AI-detection tools that also don't work. The return of the blue book and the handwritten in-class essay. Lockdown browsers, surveillance of the take-home exam, oral examinations for undergraduates. Every one of these is an attempt to re-impose, by enforcement, the scarcity that made the signal informative--and I have already argued why this must fail: no enforcement can restore a scarcity that the technology has dissolved. You cannot police your way back to a world where competent output is expensive.

But the failure is not the point. The spending is the point. Here is the tell, and it is worth teaching to anyone trying to read this moment: when an institution pays more and more to verify its own signal, you are watching the signal die. A trusted signal is cheap to accept--that is what trust means. Escalating verification cost is what a dead signal looks like from the inside.

The Snap

How does it end? Not by erosion. Credentialing systems, history suggests, do not degrade gracefully. They persist hollow, and then they snap.

The cleanest case on record is the Chinese imperial examination system. For centuries it was the sorting mechanism of the world's largest civilization: a ladder of exams on the Confucian classics that selected the scholar-official class. And for generations before its end, nearly everyone who mattered understood that its content had detached from its function--that mastery of the "eight-legged essay" certified nothing the state actually needed. The system continued anyway, hollow, decade after decade, because everything coordinated on it: family strategy, elite formation, the legitimacy of the state itself. Then in 1905, when the state's operative needs finally changed, an examination system thirteen centuries old was abolished essentially overnight.

That is the shape to expect. Not reform from within--the last post explained why the inside cannot reform itself; every adult in the building was installed with the template as a child. The discounting comes from outside, from the one party whose participation was always the point: the employer. And it comes quietly at first. A major company drops the degree requirement here; a hiring platform starts weighting demonstrated skills there; a sector discovers that its own internal assessment predicts performance better than the transcript ever did. Each defection makes the next one cheaper. Coordination unwinds the way it was built--invisibly, then suddenly. The credential will be honored everywhere, right up until, in one hiring cycle or three, it mostly isn't.

Where the Sorting Goes

Here is the question that matters more than the collapse: sorting does not disappear when the school's version of it fails. The operative need--some cheap, trusted way for strangers to rank the young--is permanent. It will be met. The only question is by what.

The economics of the answer are straightforward. Verification is expensive. Google can afford to run its own--a famously rigorous, multi-stage evaluation apparatus--because Google is Google. The small company cannot; it was always outsourcing its trust to the diploma. When the diploma fails, that company still needs somewhere to outsource its trust. Which means the successor to the credential is whatever centralizes the cost of verification--and we already know what that looks like, because parts of the economy that could never tolerate credential inflation built it long ago. The bar exam. The medical boards. The CPA. The actuarial exams. High-stakes, proctored, standardized, and--note this carefully--in person, with the technology left at the door.

That last feature is not incidental. It is the entire design. AI destroyed the take-home signal because AI can produce any output you can carry into the room. What AI cannot do is sit in the chair. Scarcity, once dissolved in the asynchronous world, gets restored through embodiment. The successor signals will be forms of witnessed performance: the proctored exam, the live interview, the work trial, the oral defense. The medieval university examined its students viva voce--by living voice, face to face--because it had no other way to know who knew. We automated our way right back to their problem, and we will arrive at their solution.

So my concrete prediction is this: something like an SAT for graduates. A voluntary, standardized, technology-free examination that a young person attaches to a job application the way one once attached a transcript--purchased not from the school but from an assessment body, taken in a room, under eyes.

And we do not have to speculate about whether such an ecosystem can exist, because in much of the world it already does. In Brazil, admission to university runs through a national examination, and it is entirely normal to spend a year or more after secondary school in a cursinho--a dedicated cram school--preparing for it. Korea and Japan have their own versions; China's gaokao is the largest examination on earth, with a shadow industry to match. Look at what these systems are, structurally: places where the school credential was already a weak signal, and where the sorting therefore migrated to an exam--with a preparation industry growing up around it like a city around a port. The pattern is the same across wildly different cultures, which by the logic of the last post means it is not a cultural preference. It is what selection produces when diplomas fail. The Brazilian cursinho year is my favorite detail in all of this (and not just because I lived in Brazil as an exchange student in high school and many of my Brazilian friends were deep in this system), because it is the system openly confessing what the diploma conceals: that the schooling was not the preparation. The exam is the actual gate, and school, whatever else it was, was custody with a story attached.

Two complications.

First: the successor will need its own noble lie. Any sorting mechanism, I argued last time, requires a story that translates its function into the individual's benefit--and the exam's story is already written: anyone can study for it. It will be about as true as the last story. Cram schools cost money; the cursinho industry is stratified by class, and everyone in Brazil knows it. The exam solves the operative problem--a cheap, trusted, scalable signal--and inherits the fairness gap intact. The narrative half of the successor, like the narrative half of everything, will be the manufactured half.

Second, and this one runs deeper. An old line holds that not everything that can be counted counts, and not everything that counts can be counted. Let me sharpen it for the present purpose: what we measure is what gets done--but not everything that can be measured is valuable, and not everything that is valuable can be measured. The tech-free exam is beautifully measurable. It certifies naked, unassisted competence: what you can do alone in a room. But the economy that is dismantling the diploma increasingly pays for something else--augmented competence, the capacity to direct the machine, to judge its output, to own the result. That is the agency I have argued is now the scarce input. And agency is precisely the thing a proctored multiple-choice instrument measures worst. So the successor system will be born with a mismatch at its heart: what is easiest to verify (the naked performance) is diverging from what is most valuable (the augmented one). The honest assessment of a young person in this era would have two parts--what can you do without the new machine, and what can you do with it--and as far as I can tell, no one is building that yet. Whoever does will be building the first evaluation instrument actually native to this century. Until then, we will measure what gets done, and call it what matters.

What Survives

One last turn, because the obvious reading of all this--school collapses--is, I think, wrong, and my own framework says why.

The crisis I have described kills one operative function: certification. But the last post named another, the one so basic it is almost never said aloud: custody. School is where the children are, and the entire adult economy is built on that fact. AI does not touch it. No technology touches it. And 2020 told us, in the institution's own voice, which of the two losses is intolerable: a year of failed learning was accepted; a year of failed custody was not.

So here is the strange endpoint the logic arrives at. The school survives the death of its own signal. The building persists, the buses run, the bells ring--because babysitting alone is load-bearing, and it was always the function least dependent on the story. What changes is that the sorting quietly moves out: to the exam, the interview, the witnessed trial, administered by bodies that are not schools. The credential lingers as a ritual document, honored the way we honor other retired currencies. And the question the last post ended on--what is education actually for?--finally receives its answer, not from a philosopher or a reformer, but from the structure itself, which will go on doing, openly, the one thing it was never willing to name.

That is not the revolution. It was never going to be. But for the first time in a century, the story and the function are about to be pulled far enough apart that everyone will be able to see the gap. What we choose to build in that gap--for our own children, with our own hands--is the real question, and it has never depended on the institution's permission.

The argument here builds on "Student Success (in the Age of AI)" and "What We Get Wrong About AI and Education."

The Conditions of Success

For years I've run a simple exercise with groups of educators that I call "the conditions of learning." I ask them to remember a time they were really learning. Not studying, not preparing for a test. A time they felt genuinely engaged, when they could feel themselves growing. I usually have to add "inside or outside of school," because in a room full of teachers people assume I want an academic answer. Some end up being from school, but most usually are not.

The stories come easily, and people love telling them, because they are so completely their own. They range across everything: a gymnastics coach, learning to become really good at a sport, a teacher who taught them to really think or write. And the stories arrive full of feeling. 

Then I ask the second question, the one that does the real work. What were the conditions that made that possible? And here is the striking part. However different the stories are, the answers converge on the same universal list. 

Someone took time. Someone believed I could do it before I believed it. Someone trusted me with something real. Someone challenged me to do something really hard. Someone knew the material deeply and could see exactly what I needed next.

The people in the room feel something shift when they identify these conditions, and the shift is both awkward and revelatory, because the conditions they just named are close to the opposite of how their own classrooms and schools are built.

That exercise has stood on its own for a long time. But the longer I sat with it, the more I realized I had been looking at a small piece of something much larger. The conditions of learning are real. They are also just one example. The same question, asked of anything you actually want, opens the same door. Not how do I get the result, but what are the conditions that produce it? What are the conditions of success?

The moment you ask it that way, something quietly changes, because most of us spend most of our energy on the result itself. We want the outcome. We rehearse having it. We get impatient for it. And the wanting does almost nothing, because results don't come from wanting. They come from conditions.

The oldest name for this is the law of the harvest. You cannot wish a crop out of the ground. You cannot demand it, or hurry it, or skip ahead to the part where the barn is full. You prepare the soil. You plant. You water. You wait through a season that does not care how ready you feel. The harvest answers to the conditions and to nothing else. It has never once answered to how badly someone wanted it.

Here is why that question does so much.

We narrate results. We do not usually narrate conditions. Almost everyone carries a vivid, detailed, emotionally rich picture of the outcome they want and almost no picture at all of what produces it. We can describe the harvest in loving detail. We go blank on the soil and preparation. They are just not as exciting. There is a gap between the story we tell about what we want and the operative reality of what would actually produce it, and that gap is where most of our wanting quietly languishes.

And much of the time we do something worse than narrate. We perform. A great deal of what we do is not building the conditions and not even honestly wanting the result. It is producing the appearance of effort and result, the signals that will read to other people as achievement. This is so ordinary that we hardly see it, because we live so far from any practical reality that would catch us out. A farmer cannot perform a harvest. The field is fed or it is not, and the barn in winter tells the truth either way. Almost everywhere else in modern life you can perform the outcome for years and never once be tested against an empty barn. That is part of why I keep coming back to the harvest. It is one of the few things left that cannot be faked.

A lot of my work is about finding that gap in order to see clearly. The space between what an institution says it does and what it really does. The space between the story a person tells about themselves and the function that story is performing. Find the gap and you find the truth. But the conditions question runs the same move in reverse. Instead of finding the gap to expose it, you find it inside your own life in order to close it. You stop living in the story of the result and start working in the reality of the conditions.

I call this productive alignment. It is what happens when you stop organizing your life or work around what you wish were true and start organizing it around what is. The wanting was never the problem. The wanting just isn't load-bearing. It feels like progress because picturing the harvest pays out a small emotional reward of its own, which is exactly why it is so easy to mistake for the work. The conditions question is the discipline that takes your attention off the result you cannot control and puts it on the soil and preparation you can.

The question forces a quieter honesty too, because the first thing it exposes is whether you even know what you are planting. It is possible to spend years cultivating someone else's harvest, a result you were handed rather than one you chose, and to mistake the exhaustion for effort. The conditions of success begin with the harvest you actually want. Get that wrong and you can do everything right and end up with a full barn that means nothing to you.

And here the first exercise comes back, because a person is not a field. Soil does not need to feel supported in order to grow corn. You do. The conditions that let a human being do hard work across a long season are the same ones those educators described: being trusted, being challenged, being encouraged, being treated as someone with their own agency. That is why the conditions of learning turned out to be a subset and not a detour. Sustained learning is one of the conditions of nearly every harvest worth having, and learning has its own conditions underneath it. It is conditions all the way down.

None of this makes the harvest certain. You can prepare the soil, plant well, tend faithfully, and still lose a season to weather you never saw coming. Bad soil you inherited. A blight that arrives without asking. A year that simply turns out hard. Anyone who has actually grown anything or worked on something substantive knows this, and knows it does not make the planting and preparing foolish. The farmer who waits for a guaranteed season never eats. The conditions were never a promise of the result. They were the only part of the result that was ever yours to build.

You cannot want a harvest. You can build the conditions for one, and then go to work.

Why School Is the Same Everywhere, and the Revolution Never Comes

Walk into a classroom in Helsinki, Lagos, Seoul, or Iowa, and you will recognize the room.

Children sorted by birth year. One adult at the front. The day cut into periods, knowledge cut into subjects, performance cut into grades. Attendance compelled by law. A credential waiting at the end, and everything before it organized around getting there.

The countries these classrooms sit in share very little. They have different languages, different religions, different economies, different ideas about the state, the family, and the good life. Over the past century they have included liberal democracies, communist dictatorships, theocracies, and military juntas--even political systems that went to war over their disagreements. And every one of them built the same room.

That should strike us as strange. If schooling were fundamentally about learning, its form would vary with beliefs about learning, and beliefs about learning actually do vary enormously, across cultures and within them. Instead, the driving variation is the story about schools. In one country, school exists to unlock individual potential. In another, to forge the socialist citizen. In another, to build the nation, or form the faithful. The narratives are incompatible with each other, but the structure of schools underneath them is identical.

When the story is interchangeable, and the structure is invariant, something interesting is going on.

What the Scholars Found

I'm not the first to notice the invariance. The best minds in education research have been circling it for decades — and what they found is worth taking seriously, because it makes the mystery deeper instead of shallower.

David Tyack and Larry Cuban, the respected historians of American school reform, named the phenomenon the "grammar of schooling": the deep organizational structures--age-grading, self-contained classrooms, subjects, periods, grades, credits--that persist across every wave of reform. Their history shows a century of serious, well-funded, passionately led attempts to change that grammar: the Dalton Plan, the Eight-Year Study, open classrooms, and flexible scheduling. Every one either failed outright or was absorbed. The school kept its structure and incorporated a diluted, domesticated version of the innovation. Their famous conclusion: schools change reforms as much as reforms change schools.

Cuban ran the technology-specific version of the study in Teachers and Machines. Film was going to revolutionize the classroom. Then radio. Then television. Each was announced as transformation; each was assimilated at the margins; the room remained the room.

Meanwhile, the sociologists John Meyer and Francisco Ramirez documented the global side of the puzzle: mass schooling spread to virtually every nation on earth as a single institutional model. It was often adopted whole, sometimes before a country had an economy that could employ its graduates. Nations did not converge on this model by independently discovering that it worked. They adopted it the way one adopts a uniform.

So the pattern is documented, exhaustively, from two directions: the model resists change from within and replicates without limit across the world. What's been missing is the explanation.

Tyack and Cuban's answer was that the public--parents, employers, students themselves--carries a "cultural template" of what a real school looks like, and departures from it read as illegitimate regardless of learning outcomes. Teachers who tried radical structures found parents demanding grades and subjects back. Meyer and Ramirez's answer was that nations adopt the model to gain "legitimacy."

These are observations posing as explanations. Why does the public carry the template so tenaciously that parents will fight to restore grades even when their children are learning more without them? Why does this particular model confer legitimacy on a state? 

The best scholarship in the field documents the grip and then stops, one level short of the machinery. I want to describe the machinery.

The Metaphor Is the Mechanism

Let's start with Tyack and Cuban's own word: grammar. They chose it because grammatical rules have a peculiar character. You acquired the grammar of your native language in childhood without anyone teaching it to you explicitly. It runs below your awareness. You cannot state most of its rules, yet you follow them perfectly, and when someone violates them you don't reason your way to an objection, because the sentence simply feels wrong.

They meant this as an apt metaphor. I think it's a literal description of the mechanism.

Human beings come equipped with a subconscious learning system that I call the adaptive mind, whose job in childhood is to figure out what this particular social environment requires for belonging, and to install those requirements so deeply that they stop feeling like lessons and start feeling like reality. This system doesn't evaluate what it installs. It can't distinguish between survival programming and selfhood. Whatever the environment consistently presents during the developmental window gets written in as the way things are, and deviations from it later trigger not disagreement but alarm, the felt sense of wrongness that needs no argument.

Now consider what school is, from the adaptive mind's point of view. It is roughly twelve years of daily immersion, for essentially every child, during precisely the window when this installation runs hottest. Whatever else school teaches, the one lesson it delivers to every student without exception--more reliably than reading, more durably than mathematics--is what school is. The sorted ages, the adult at the front, the subjects, the bells, the grades: these are not experienced as design choices that could have been otherwise. They are installed as the definition of the thing itself.

This is why the template grips. The parents who demand grades back are not weighing evidence about learning outcomes and getting the answer wrong. They are running programming installed in them, as children, by the very institution they are now defending. A gradeless classroom registers the way an ungrammatical sentence does, which is not as a different choice but as a violation, felt before it is thought.

And this is why the grammar survives every reform. The institution manufactures its own enforcement constituency, one generation at a time. Every adult who might change the school was formed by the school. Reformers are asking people to redesign a structure their own nervous systems experience as the natural order, and asking them to do it on behalf of children whose nervous systems are absorbing that same structure right now. Each new generation arrives with no inherited immunity, gets the installation, and grows up to defend it. The school is not merely resistant to change. It is self-replicating.

The Story That Runs the Machine

But installation alone doesn't explain the devotion. For that, let's look at the story the institution tells, because it is, structurally, a masterpiece.

Mass schooling operates at the level of populations: it sorts the young, certifies them, occupies them, and habituates them to institutional life at a scale no other invention has matched. But it is never sold at the level of populations. It is sold to each family, one at a time, as private advantage: this is how your child rises.

Consider what that promise attaches to. Little in a parent's psychology runs deeper than the survival prospects of their own child. By routing its story through that channel, the institution recruits the most powerful motivational machinery humans possess, and points it at enforcement. A parent defending the grammar isn't doing the state's work; as far as their own nervous system is concerned, they are protecting their child. Deviation from the template (homeschooling, alternative school, or skipping the credential) doesn't feel like an educational experiment. It feels like gambling with your child's life.

This is why the system requires so little visible coercion. Provide a compulsory-attendance floor, a monopoly on the credential, and the story of individual benefit--and the managed population supplies the management energy itself, voluntarily, fiercely, at every kitchen table.

None of this is a modern discovery. The first systematic blueprint for state education in the Western canon, Plato's Republic, is built on exactly this pairing. Plato's city needs its young sorted: gold, silver, bronze, each to their station. And Plato, unsentimental about what sorting requires, prescribes the story to go with it--"the noble lie": citizens are to be told that their metal was placed in their souls by the god, so that each accepts his sorted position not as imposition but as identity. Plato understood something his modern successors rarely say aloud: a sorting institution cannot run on truth. It requires a myth that translates the institution's function into the individual's nature, a story in which being sorted is the benefit.

Modern schooling did not descend from Plato; it rediscovered him. No committee ever voted to implement the noble lie. But any institution that sorts the young at scale faces the structural problem Plato solved on paper, and the solutions converge. Ours is the story of merit and opportunity: your grades reveal what you are, your credential certifies it. This does the work of the myth of metals precisely, with one improvement Plato couldn't have imagined: his guardians had to knowingly tell the lie, while ours believe it, having been installed with it as children themselves.

Why It Conquered the World

None of this required a designer, and this is the part I most want to be understood, because the argument is easily mistaken for a conspiracy theory and it is precisely the opposite of one.

The model we now find everywhere took recognizable shape in Prussia, and it spread the way successful adaptations spread: by selection. States that adopted mass schooling gained things that mattered in the competition between states: soldiers who could follow written orders, workers habituated to the clock and the bell, administrators, and a population rendered legible and manageable. Those states outperformed their neighbors, and their neighbors copied them, the way you copy a winner. No one needed to intend the outcome. The institutional form that most effectively solved the problem of managing large young populations, while telling the most compelling story about individual benefit, outcompeted every alternative form, and its descendants now cover the earth.

This is a general law of cultural evolution, and schooling is only its most complete example: the cultural forms that survive and spread are the ones that most effectively work with, and on, evolved human psychology. This is regardless of whether they are true, and regardless of whether they serve the humans running them. Ubiquity is not evidence that an institution is right. It is evidence that it is fit. The question worth asking of any universal institution is never "why is it so successful?" but "what, exactly, is it fit for?"

Two honest concessions are needed. First, the story of individual benefit is not a pure fiction. Mass schooling delivered real literacy and numeracy to whole populations, pulled children out of factories, and produces genuine mobility for some (and it must, because a story with no true positives stops being believed). The claim is not that school delivers nothing. It is that the form is optimized for management, and the delivery is calibrated to whatever keeps the story load-bearing. Second, the grammar also solves a real engineering problem: one adult, thirty children, at a price the public will pay. The grammar persists because it is installed, defended, and cheap. These reinforce each other; they don't compete.

The Day the Function Spoke

If you want proof that the operative functions, not the story, run the institution, you don't need a century of reform history. You only need to remember 2020.

Among schooling's functions, one is so basic it is almost never named: custody (or bluntly, babysitting). School is where the children are for six hours a day, five days a week, nine months a year. The entire adult economy is built on that fact. Every workplace schedule silently assumes it. This is one of the quiet reasons the grammar cannot change: proposals that touch the school day or calendar collide not with pedagogy but with the labor force.

Then the schools closed, and for a moment the institution's real priority order became visible. The tolerable loss, it turned out, was the learning: a year of remote instruction that almost no one believed was working was accepted. The intolerable loss was the custody. Parents couldn't work, and the economy said so immediately, out loud, in those words: schools had to reopen for the economy. It may be the only time in living memory that the operative function has spoken in its own voice, in public, without the story. Then the doors opened, the language of learning and flourishing resumed, and we seem to have agreed not to mention it again.

The Revolution That Never Comes

Which brings us to the claim many of us have heard over decades, and are hearing again right now: this is the technology that will finally revolutionize education.

It was said about film. About radio. About television, computers, the internet. Each time, the revolution was announced; each time, the school absorbed the technology and basically remained itself. Cuban documented this pattern decades ago without quite explaining it. Now we can: the revolution is always promised to the story--individual flourishing, unlocked potential--and stories don't run schools. The operative functions do: sorting, credentialing, custody, the management of the young at scale. A new technology threatens none of these, so none of these change. The story flexes just enough to include the new machine, and the machine is domesticated. Schools change reforms more than reforms change schools.

So no, AI will not revolutionize education. A century of experience says otherwise.

But this time something different is happening underneath. The bargain at the center of schooling has always been compliance for credentials: produce the assignments, sit the years, and receive a signal employers trust. That signal worked because compliant output was costly, and it took a childhood to produce. AI makes compliant output free. The story will survive this, as it survives everything, but the bargain may not.

What's coming isn't a revolution in education. It's likely to be a crisis in what education is actually for, since the two have never been the same thing.

That's a longer argument, which I've made in part here and here.

Media Literacy Without Human Literacy: Narrative Enforcers Dressed as Critical Thinkers

In classrooms, libraries, and professional development sessions across the country, media literacy is now presented as an essential defense against misinformation, propaganda, and manipulation. Students learn to check sources, identify bias, spot emotional appeals, and verify claims. Teachers and curriculum designers position these programs as vital preparation for democratic citizenship in a complex information environment.

Yet a deeper examination reveals a persistent and troubling pattern. Most media literacy education operates at a strikingly superficial level. It equips people with procedural skills while leaving untouched the underlying architecture of human psychology that makes sophisticated manipulation possible in the first place. The result is not a population better able to perceive reality, but a class of more sophisticated narrative enforcers—individuals who have internalized the boundaries of acceptable discourse and now police them with the language of critical thinking.

This is not a failure of good intentions. It is a structural outcome. Media literacy programs, as currently designed and delivered, are largely not human literacy competent. They lack the foundational understanding of why human beings are so reliably vulnerable to narrative exploitation. Without that foundation—what I call human literacy—these programs inevitably drift into complicity with the very systems they claim to critique.

The Surface Curriculum and Its Hidden Function

Standard media literacy instruction typically includes:

• Distinguishing news from opinion or sponsored content
• Evaluating domain authority and source credibility
• Recognizing loaded language, images, and emotional manipulation
• Fact-checking claims against established institutions
• Understanding algorithms, echo chambers, and filter bubbles

These skills have limited value. They can help individuals avoid the most obvious fabrications. But they fundamentally misdiagnose the problem. They treat manipulation as primarily a matter of individual error or malicious outsider fabrication. They rarely ask the structural questions: Why do certain narratives persist across institutions despite contradictory evidence? What incentives shape what counts as a "reputable source"? How does our evolved psychology make us active participants in our own manipulation?

This is the narrative-operative gap in media literacy education itself. The idealized narrative is empowerment through critical thinking. The operative function is often quite different: training students to defer to institutional consensus, to treat credentialed authority as the measure of truth, and to experience dissent from dominant narratives as a form of personal or cognitive failure. The programs perform skepticism while installing compliance.

This dynamic mirrors the broader pattern I have written about elsewhere. Institutional education frequently functions as a delivery system for adaptive mind programming—the installation of local consensus as a survival imperative. Media literacy, in this context, becomes an advanced module. Students learn to perform the appearance of independent analysis while the deeper mechanisms of exploitation remain invisible and unexamined.

The Exploit, Blame, Shame Mechanism in Information Systems

The Law of Inevitable Exploitation predicts that systems and behaviors which most effectively harness available resources—including evolved human psychology—will survive and spread, regardless of their relationship to objective truth or human well-being. Media and information ecosystems are no exception.

When large-scale manipulation occurs through mainstream institutions, the cultural response rarely involves structural examination of those institutions. Instead, we see the familiar three-stage pattern:

1. Exploit: Psychological vulnerabilities are leveraged at scale through narrative management, selective framing, and coordinated messaging.
2. Blame: Individuals are held responsible for "falling for" the resulting beliefs or behaviors.
3. Shame: Those who accurately perceive the manipulation are pathologized as conspiracy-minded, cynical, or lacking media literacy.

Media literacy programs, by focusing exclusively on individual skills rather than the architecture of exploitation, participate in this mechanism. They become part of the enforcement layer—teaching people to blame themselves and others for outcomes that are structurally produced. This is structural victim blaming in educational form.

The intensity clue is often visible here. Emotional defensiveness around certain topics, or the quick labeling of structural questions as "conspiracy thinking," frequently signals that a load-bearing narrative is being protected. Genuine media literacy would treat that intensity as diagnostic information rather than as evidence that the questioner has failed.

AI Acceleration and the Rush to Superficial Expertise

The emergence of generative AI has intensified this problem rather than resolving it. Suddenly, the same educators and organizations that taught surface-level media literacy are repositioning themselves as experts in "AI literacy." They offer workshops on prompt engineering, detecting AI-generated content, understanding algorithmic bias, and using AI "responsibly."

What is almost entirely absent from these efforts is any engagement with the deeper questions the technology raises:

• How does AI interact with the Separated Mind Architecture—the gap between conscious intention and the subconscious heuristics shaped by Paleolithic survival pressures?
• In what ways does prolonged interaction with AI systems produce algorithmic capture and model capture, subtly reshaping users' thinking, writing, and perception over time?
• Why do institutions that have demonstrated little capacity to perceive manipulation in human media systems suddenly claim authority over AI ethics and governance?
• How might AI be used not merely to automate existing surface analysis, but to perform the kind of scaled pattern recognition across human records that reveals structural regularities invisible to any single human observer?

The pattern is predictable. Those operating within institutional adaptive mind programming rush to claim expertise in the new domain without having developed the metacognitive distance required to see the previous domain clearly. They do not recognize the Returning Prisoner's Dilemma because they have not undergone the disorientation of genuine perception. They have not cultivated the outsider's perspective that provides analytical access to the architecture of capture and exploitation.

As a result, they systematically marginalize or fail to engage with thinking operating at the structural level—thinking that examines evolutionary mismatch, functional fictions, coalitional psychology, the fractal nature of exploitation, and the gap between idealized narratives and operative functions. These analyses threaten the consensus that surface media literacy is designed to protect. The limitation is not a moral failing; it is architectural. The adaptive mind treats challenges to installed consensus as existential threats.

What Human Literacy Actually Requires (the Steve Hargadon Version)

If media literacy is to serve human flourishing rather than institutional narrative management, it must be grounded in human literacy—an understanding of the cognitive architecture that makes us susceptible to manipulation.

This foundation includes recognizing several structural realities:

The Separated Mind Architecture. Human cognition operates in layers with limited direct communication. The adapted mind (evolutionary firmware) runs ancient survival heuristics optimized for small-group Paleolithic environments. The adaptive mind (cultural software) rapidly installs the specific performances required for belonging in one's local environment. Consciousness (the rider) makes real decisions but from a menu it did not design. Narrative is the primary bridge between layers—and therefore the primary vector for both genuine understanding and sophisticated manipulation.

The Paleolithic Paradox. Our psychological machinery was forged for environments radically different from the ones we now inhabit. Status-monitoring, coalition-detection, authority deference, and approval-seeking operate continuously, often producing anxiety, depression, and complicity that feel personal but are structurally generated. Most behavior labeled "self-sabotage" in information consumption is actually real sabotage—external systems exploiting these heuristics more effectively than we understand them.

The Narrative-Operative Gap as Diagnostic Tool. Every human system—individuals, institutions, civilizations—maintains an idealized public narrative and an operative reality. The gap between them is not hypocrisy but architecture. Identifying the gap reveals operative truth. Media literacy without the capacity to perceive and analyze this gap at scale is not literacy; it is sophisticated performance within the gap.

The Law of Inevitable Exploitation. Systems that most effectively exploit the psychology they encounter will tend to survive and spread. This is not a conspiracy claim but a structural prediction. Conspiracies of coordination exist on a continuum with emergent coalitional dynamics; both are made possible by the same underlying architecture. Distinguishing between them requires structural analysis, not reflexive dismissal.

Genuine media literacy built on this foundation would teach students to ask different questions:

• What is the idealized narrative of this institution, story, or technology, and what is its actual operative function?
• Whose evolved psychology is being exploited here, and through what mechanisms?
• What load-bearing fictions must be maintained for this system to continue operating?
• How does my own adaptive programming make me vulnerable to this particular form of manipulation?
• What structural constraints would be required to close the narrative-operative gap, rather than merely teaching individuals to navigate it more skillfully?

These questions are harder. They require confronting the possibility that many of our most trusted institutions operate with significant gaps between stated mission and actual function. They require developing the capacity to observe one's own adaptive mind programming rather than merely performing within it. They require accepting that dissent from consensus is not automatically evidence of error, and that consensus itself can be a measure of social pressure rather than truth.

The Cost of Superficial Competence

The tragedy of current media literacy efforts is not that they teach nothing. It is that they succeed in teaching the wrong thing—or rather, in teaching skills that primarily serve institutional survival and narrative coherence rather than human perceptual capacity.

In an age of AI, where the ability to generate, reinforce, and personalize sophisticated narratives is scaling exponentially, this gap becomes existentially dangerous. We are producing populations skilled at identifying crude fakes while remaining largely blind to the more elegant and institutionally embedded forms of manipulation. We are training people to enforce boundaries they did not set and cannot see.

What we need is not more media literacy layered on top of unexamined psychology. We need human literacy as the foundation—understanding the architecture of our own minds and the systems that have evolved to exploit it. Only from that foundation can media literacy become something other than sophisticated complicity.

The alternative is to continue producing graduates who can competently police the shadows while remaining unable to perceive the machinery casting them. That is not education. It is the next evolution of narrative enforcement, now wearing the respectable clothing of critical thinking and digital citizenship.

The Cost of Pretending

Every adult knows a particular kind of tiredness. It is not the tiredness of hard labor or too little sleep. It is the tiredness of maintenance: the low, background exhaustion of pretending things that you no longer believe or that are just not true.

The earliest training for many of us happened at Christmas. We discovered that Santa Claus is a fiction. But instead of exiting the fiction, we are recruited into maintaining it. We learn to perform belief for younger siblings or friends, to nod along at the dinner table, to protect, and even to revel in, the story. That moment, when we realize the gap between what is said and what is true, and are taught to bridge it with our own behavior, is the doorway into adult cognitive life.

Santa is our first inculcation into the structure. Every system that asks you to participate in knowing lies is running the same play.

The Load

There is no widely used name for what I want to describe. Cognitive dissonance names the discomfort of holding contradictions, but it emphasizes the urge to resolve them. The cost of pretending is different: it is the continuous energy required to sustain the contradiction because resolving it would mean exiting the system that requires it. Doublethink describes the capacity to hold two beliefs at once, but it is usually framed as something totalitarian and extreme, not as a feature of ordinary life. Emotional labor and impression management capture pieces of the work, but they stay at the level of single interactions or service jobs.

What is missing is an integrated term for the cumulative cognitive and emotional cost imposed on members of any system whose stated purposes or idealized narratives diverge from its operative functions. That cost includes the work of performing and performing the stated or idealized purpose, the work of operating in the actual one, the work of defending the narrative when challenged, and the work of suppressing your own awareness of the gap when it threatens to surface.

The reality is that most of us are running maintenance on dozens of fictions simultaneously, and the aggregate load is enormous. Because the maintenance has been running since childhood, we do not experience it as a load. It is automatic, like breathing. It shows up instead as a baseline fatigue that has no obvious source. It is the exhaustion of being an adult, the sense that simply being awake requires more energy than it once did. Some of that is overstimulation and information density. Much of it, I would argue, is the cost of pretending, and the maintenance of it becomes identity itself.

Cognitive Capture

The load does more than tire. It disables.

Milton Erickson found that deliberate confusion could induce suggestibility by overloading conscious processing. The gap load operates as a chronic, distributed version of the same effect. The maintenance work—running every potential statement through the mental arithmetic of what this audience can tolerate—occupies the working memory that would otherwise be devoted to evaluating the system itself. The result is a population too cognitively busy to dissent, not because they fear punishment, but because their executive function is already fully deployed keeping the fiction airborne. The gap is stable not merely because exit is costly, but because maintaining the gap consumes the cognitive capacity required to plan an exit.

This is why modern institutions and organizations tend to succeed and grow. They capture human time and labor not despite the confusion they generate, but because of it. Debilitation is the point. The employee is not merely obedient; he is too cognitively occupied to formulate alternatives. The fog is structural. The institution expands because it produces exactly the cognitive state that prevents resistance. And temporary relief can be provided through distraction and entertainment. 

Pre-Adaptation

This is not accidental architecture. The modern classroom descends directly from the Prussian system of the early nineteenth century, imported by American reformers like Horace Mann after visits to Europe. The Prussian model was not designed primarily to cultivate independent thought; it was designed to produce controllable citizens and soldiers through regimented habit, replacing localized apprenticeship with centralized, age-segregated drill. The gap between its stated purpose (enlightenment, education, preparation) and its operative function (obedience, punctuality, submission to arbitrary authority) was present at the founding. The child learns to sit still, to speak only when permitted, to shift attention at the bell, and above all to treat the teacher’s frame as paramount, regardless of his own judgment. These are not pedagogical techniques; they are behavioral protocols that consume working memory with compliance tasks.

In this light, the schoolroom is the first and most intensive training ground for the cost of pretending. The student must simultaneously track what they actually understand, what the test requires them to say, and what social cost attaches to either. They must express gratitude for instruction they recognize as irrelevant, and show enthusiasm for evaluations they know to be arbitrary. The aggregate cognitive load is enormous precisely because it is continuous and pre-emptive: every potential utterance is filtered through the gap before it is spoken. By the time they reach adulthood, the skill is automatic. They have been pre-adapted to maintain institutional fictions not because they believe them, but because they have never known a social environment that did not require it.

The Inventory and Its Symptoms

If you inventory the major domains, the breadth becomes visible. Patently false but agreed-upon political narratives, economic narratives, religious narratives, romantic narratives, familial ones, professional, educational, medical, journalistic, interpersonal. Each has its own maintenance requirements. Each contributes to the aggregate. Most of us have never listed the fictions we are actively maintaining, which is part of why the load remains invisible.

This manifests in predictable ways. There is the baseline fatigue. There are the momentary lapses, like the unexpected honesty in a conversation that feels like relief even when the content is painful, or the hidden-camera interviews that uncover candid admissions of lying. There is irritability that flares when a fiction is challenged, because the challenge threatens to increase maintenance costs. And there are the breakdowns: burnout, alienation, depressions with no obvious cause. When pretense maintenance exceeds capacity, the symptom is often emotional volatility. The instability is sometimes diagnosed as a mental health problem, which it functionally is, but the underlying source is the gap between narrative and reality becoming too expensive to bridge.

What the Load Explains

Once you name the cost of pretending, several patterns snap into focus.

Social intelligence is, at bottom, the efficient management of pretense. The socially intelligent person is not merely "good with people." They are tracking multiple simultaneous fictions and operating within them without breaking any. They know which version of the story to tell in which room. They know when honesty is costly and steer clear of it. This is real skill. It is also, structurally, gap-load efficiency. The people we call socially awkward are often not deficient; they are simply declining to pay a cost that others pay automatically. They pay a different cost—social exclusion—while the neurotypical mainstream pays in continuous, invisible maintenance.

Career sorting is partly pretense-load selection. Roles with low maintenance demands attract people who cannot or will not do the work: tradespeople, engineers, certain scientists and artists. Roles with high maintenance demands—management, sales, politics, law, public relations—attract people who carry heavy loads easily, and they pay a premium for it because the labor is real. Personnel patterns that look like "personality fit" are often better explained as load-tolerance sorting.

Truth-telling tolerance is measurable across cultures. Look at the range of topics permissible in polite conversation. Look at how a culture treats its truth-tellers. Cultures with high truth-telling tolerance have lower gap loads because the gap is partially acknowledged. Cultures with low tolerance impose higher loads because more maintenance is required. When a culture moves simultaneously toward lower tolerance across multiple domains, you should expect a rising baseline of cultural exhaustion that is hard to name.

Fundamentalism, religious or secular, offers a specific relief. The fundamentalist does not have to negotiate between stated and operative because, in their world, the stated is the operative. The maintenance cost collapses to near zero. The relief is real. The rigidity is the price. People will accept significant rigidity in exchange for the rest this provides, which explains why fully believed positions are surprisingly stable over time. They are restful. The most exhausting position is always the middle: half-believing, participating without conviction, running the narrative while remaining aware that it is a narrative. Full exit is freedom; full belief is rest; the middle is cognitive purgatory.

Therapy, when it works, is often the structured dropping of pretense. The patient arrives with symptoms that are the byproduct of maintenance work that has become unsustainable. The therapist’s job is to hold a space where the gap can be acknowledged: the marriage is not what it appears, the job is not what the story says, the family mythology is not the history. Each acknowledgment is a piece of maintenance work set down. Bad therapy keeps the fictions intact and merely adds a credentialed participant to the pretense. Good therapy is permission to stop pretending in one carefully bounded room.

Emotional intensity itself becomes legible through this frame. Outrage in defense of a narrative is not evidence of sincerity; it is evidence of maintenance strain. The intensity tracks how much the fiction is doing for the defender. It is also fuel: revivals, movements, obsessions, and cults use intensity to power the maintenance that would otherwise collapse. Emotional volatility with no clear trigger may just be the cost of pretending exceeding the payer’s capacity.

Alignment as Relief

The framework produces a clean prediction: the relief of alignment is the relief of stopped maintenance. A well-aligned system is one in which members do not have to expend cognitive effort to hold its official story together. A poorly aligned system taxes its members continuously just to keep them participating.

That means honest spaces are genuinely restorative. They are places where the gap is acknowledged, and the maintenance work pauses. The relief you feel in a conversation where the pretenses drop is like energy returning to a muscle that has been clenched for years. I have a theory that businesses and organizations that minimize this gap, which are good at aligning their stated narrative with their actual functional activity, are in a form of strategic alignment that produces tangibly positive results.

Naming the cost does not eliminate it, but it changes the relationship to it. Most people experience the fatigue as a private failure—a lack of resilience, a character flaw, a need for more sleep, or better boundaries. Recognizing it as a structural tax imposed by the gap between narrative and function reframes the exhaustion as a real cost paid to real systems. And once you have the name, you can begin to inventory the specific fictions that are costing you the most.

You can ask which ones you are willing to keep maintaining, which ones you can exit, and which systems you might be able to align rather than endure. The cost of pretending is already being paid. The only question is whether we will notice the bill.

Student Success (in the Age of AI)

What the New Machine Can't Supply

A student sits down to write the essay her teacher assigned. She opens a chatbot, pastes the prompt, and nine seconds later, she is holding a competent five-paragraph response. She changes two sentences so it sounds like her and submits it. Her teacher, facing sixty of these and no more hours in the day than anyone else, runs each one through the same kind of machine to generate constructive feedback, pastes it into the margins, and assigns the grade. A few seconds of generation on one end, a few more on the other, and the circuit closes with no human thought anywhere inside it. She earns an A. Nothing about this is remarkable anymore. What is remarkable is what it means. The grade measures nothing. The feedback taught nothing, because no one read the essay, and no one wrote the response. The assignment existed to test whether she could produce the output; she produced the output; she learned nothing; the teacher taught nothing; and the system has no way to tell the difference, because from the system's point of view nothing is wrong. Every box was filled. Two machines spoke to each other, and two humans stood at the endpoints holding the receipts.

This is not a story about cheating. It is a story about a bargain coming apart. The bargain was so old and so deep that most of us mistook it for the natural order of things. For two centuries, school sold a deal to every student who walked through its doors. The deal was: comply, perform, accumulate the credential, and the credential will convert into success. Do the work, get the grade, get the diploma, get the seat at the university, get the job. The new machine has just reached into the middle of that sentence and the first thing it did was to automate the part the student was supposed to supply. And in doing so it has revealed something the bargain was carefully built to keep hidden: that the thing school actually rewarded was never learning. It was compliance. The two had simply been close enough, for long enough, that almost no one needed to tell them apart.

I want to argue something that sounds, at first, like a paradox and turns out to be the most practical claim I know how to make. In the age of artificial intelligence, agency is no longer one path to success among many. It is the only thing left that can actually produce it. Not because agency is noble, though it is, and not because self-direction is a nicer way to raise a human, though it is that too. Agency has become load-bearing for a more pragmatic and structural reason: it is the one input the new machine cannot supply, cannot fake, cannot simulate, and cannot replace. Everything else it can now do. That single fact rearranges the entire landscape of what it means to learn, to teach, and to succeed.

The Bargain, and Why It Worked

To see why the bargain is breaking, you have to see why it held in the first place, and the honest answer is uncomfortable.

Every institution runs on two stories at once. There is the story it tells about itself, the aspirational one printed on the mission statement, and there is the thing it actually does, the operative function that pays its bills and reproduces it year after year. Schooling's official story is the development of the individual mind. Its operative function has been sorting. School took a population of children and ranked them, stamped them, and delivered them in order to the next stage of the economy. It did this by exploiting something real in human wiring: our deep, ancient deference to authority, our hunger for approval, and our compulsion to monitor our standing relative to everyone around us. Put a child in a room, attach a grade to their performance, and the evolved system does the rest. They will compete for the grade. They will internalize the grade. They will, in time, mistake the grade for themselves.

I have spent years writing about the mechanics of this under a name I keep returning to: the Game of School. The game has rules that are not only unstated but also often invisible to those most affected by them. The rules are not about learning; they are about reading the teacher, supplying the expected answer, managing the appearance of effort, and never confusing the performance with the thing it pretends to measure. The students who thrive are not the ones who learn the most, but the ones who decode the game fastest — those who grasp early that the assignment is a transaction, that understanding is optional, and that the reward goes to the one who delivers the output the institution wants to see. The cruelest part of the game is that it teaches most students that they are not good learners. It pronounces a personal deficiency, a verdict on the child rather than the design. A structure built to rank will always produce a ranked bottom, and then it will tell the bottom that the ranking was about them.

For two hundred years, this was, in the coldest sense, practical. The economy on the other side of the schoolhouse door wanted exactly what the game produced: people who would show up, follow instructions, tolerate boredom, defer to authority, and finish assigned tasks whether or not they cared about them. School was a remarkably efficient training apparatus for an industrial order that ran on compliant labor, and its genius was that it disguised training as development and conditioning as growth. The cover story (we are here to cultivate your mind) let everyone participate in the operative function (we are here to sort and shape you for your station) without ever having to say it. The gap between the two stories was wide, but it was stable, because the credential at the end carried enough real signal to keep the whole arrangement productive.

What made this defensible, what kept the gap between the two stories from becoming intolerable, was that the credential did carry real information. A diploma, a grade, a degree. These worked as signals because the compliance they certified was expensive. Someone had to actually sit there and do the reading, grind through the problem sets, produce the essay, and show up for years. The cost of the performance is what made it mean something. It correlated, imperfectly but well enough, with the traits an employer wanted: that you would persist, follow instructions, finish what you started, and defer when required. The grade was never a measure of learning. It was a measure of trainability, and trainability was valuable, and so the fiction was functional. Everyone could pretend the credential meant understanding because it at least meant something, and that something was useful.

The bargain worked not because it was true but because its central mechanism — costly, human, effortful compliance — was scarce. The whole edifice of grades, admissions, and credentials was an instrument for measuring a scarce thing. Take away the scarcity, and the instruments measure nothing.

What the New Machine Severs

That is precisely what has happened, and it has happened faster than any institution was built to absorb.

When a model can produce the compliant output — the essay, the lab report, the problem set, the code, the cover letter — in seconds and at no cost, the performance stops being expensive. And the moment the performance stops being expensive, it stops being a signal. The grade decouples from the trait it was quietly standing in for. The diploma certifies that a student had access to a chatbot, which is to say it certifies nothing at all. This is not a problem that better testing or cleverer plagiarism detection will solve, because it is not really a problem of dishonesty. The signal worked because it was costly to simulate. It is now free to simulate. No enforcement can restore a scarcity that the technology has dissolved.

Notice what this does to the gap between school's two stories. For two centuries, the operative function, sorting through certified compliance, could hide behind the official one, developing the mind. Because the certified compliance was at least real. Now the operative function has been hollowed out from the inside. The new machine is what performs the compliance, so the sorting mechanism sorts noise, and the official story it was hiding behind is suddenly standing in the open with nothing underneath it. The fiction did not collapse because someone exposed it. Fictions almost never do; we are far too invested in our comfortable stories to give them up to mere argument. It collapsed because its load-bearing mechanism was automated to zero; a fiction can survive exposure, but it cannot survive the quiet removal of the thing that was actually doing the work.

And here is where this is more than a story about schools. The same severing is happening everywhere, all at once. The compliant performer in the office, the one whose value was producing the standard memo, the routine analysis, and the competent deck, is being exposed by the same blade that exposed the student. Across every domain where a human was paid to supply effortful, rule-following output, the new machine is removing the scarcity that made that output worth paying for. AI is, among other things, a great revealer. It is automating the performed-compliance layer of human work at every level of organization at the same time, and as it strips that layer away, it leaves visible the thing that was always underneath, the thing that was never the point of the credential and never could be automated: the human's capacity to direct the work rather than merely perform it.

What Is Left

So what survives? When the new machine can produce any output you can specify, what is the thing it still cannot supply?

It cannot supply the specification. It cannot decide what is worth making, or judge whether what it made is any good, or know when the brilliant-sounding answer is subtly wrong, or care about the outcome, or own the result. It cannot want anything. It can generate a thousand directions but not a single preference. The capacity to choose a direction and pursue it, to bring judgment to bear, to take responsibility for the result as yours — this is what I mean by agency, and it is the bedrock on which all genuine learning has always rested.

Let me be precise about what agency is not, because the word gets used loosely. Agency is not effort; the most diligent student in the room may have no agency at all, having only ever obeyed with vigor. Agency is not compliance; it is closer to compliance's opposite. And agency is not raw intelligence; plenty of brilliant people have outsourced every decision that mattered and never noticed. Agency is the capacity to be the author of your own action, to set the aim, to steer, to evaluate, and to own. It is the one human function that, by definition, cannot be delegated to the new machine, because the moment you delegate it, it is no longer yours. The new machine can carry out your intent. It cannot have your intent for you. Try to hand it that, and you have not gained a tool; you have literally disappeared.

This is why the old system could punish agency for two centuries and still function. In a world where compliance was scarce and valuable, the self-directed child was an inconvenience. The one who asked why, who wanted to do it differently, who would not simply perform on command, could be classified as defective. School had no use for that and often crushed it, and the economy absorbed the compliant graduates it produced, and the arrangement held. Agency was always the real substance of learning, but compliance was a good-enough proxy in a low-machine world, so we built an entire civilization-scale institution that optimized for the proxy and often treated the actual substance as a discipline problem. AI removes the proxy. For the first time, the thing school spent two centuries suppressing is the only thing with any value left.

The Choice Every Learner Now Faces

Put a powerful new machine in the hands of a person, and you have not determined anything yet. You have only sharpened a question that was always there and can now no longer be avoided. There are three things a person can do with a tool this capable, and which one they choose decides everything.

They can surrender to it: let it think in their place, accept its outputs without judgment, hand over not just the labor but the direction and the discernment. This feels like efficiency and is, in fact, erasure. The person who surrenders brings nothing the machine did not already have, and so, predictably, becomes redundant to their own life. The capacities they stop using atrophy, exactly as a muscle does, until the surrender is no longer a choice but a condition.

They can offload to it: hand over the parts of the work that do not require them, the boilerplate, the grunt labor, and the lookups, while keeping the direction and the judgment for themselves. This is roughly neutral and often good. It is what a calculator is to a mathematician: it frees attention for the part that is actually hard and actually theirs.

Or they can sharpen against it — use the machine as something to think with, a tireless interlocutor that helps them articulate, test, and refine what is theirs, while they retain authorship the entire way through. The person who sharpens does not become smaller as the tool grows more powerful. They compound. Every increase in the machine's capability is an increase in their reach, because they are still the ones steering.

The same tool, in the same hands, amplifies one person and replaces another, and the variable that determines which is not intelligence, or wealth, or access. Everyone now has access. The variable is agency. The machine is a mirror with a multiplier: it returns your own degree of self-direction, magnified. Bring agency, and you become formidable. Bring none, and you become unnecessary. This is the whole game now, and it is being played, mostly unconsciously, by every student and every worker alive.

Why Success Is Now Agency

There have always been two ways to argue for agency, and they have always seemed to pull in different directions. The instrumental argument says: cultivate agency because it is the best route to the success you already want, i.e., the grades, the admission, the career. The intrinsic argument says: forget the metrics, they were always a proxy; agency is what success was supposed to mean all along, the self-authored life being the only one worth calling successful. The first argument is persuasive to a school board and slightly cynical. The second is true to anyone who has thought hard about it, and useless for getting a program funded. For most of modern history you had to pick one, because in a world where compliance reliably produced the credential, agency and metric-success genuinely were separate. You could succeed by the numbers with no agency at all, simply by playing the game well.

AI welds the two arguments into one. In a world where the new machine performs the compliant half, the only remaining source of the metric-success everyone still wants is agency. The student who can direct, judge, and own, who can use the machine to go further than either could alone, is the one who produces work of real value. And real value is what the credentials were always trying and failing to measure. The agentic learner gets the tangible wins too, not as a happy accident but as a structural necessity, because agency has become the scarce input that the entire economy is now short of. You no longer have to choose between teaching a child to be a self-directed human and teaching them to succeed. The age of AI makes those the same instructions, where the thing that is true and the thing that is useful have stopped diverging.

Where Agency Grows

If agency is the whole game, then the only question that matters for education is how a human acquires it. This is exactly where the old institution cannot follow, because its entire method is the suppression of the thing now most needed.

You cannot manufacture agency on a factory line, for the same reason you cannot order someone to be spontaneous. The factory model of schooling works by removing choice, standardizing the path, and rewarding obedience to it. Every one of those mechanisms is the precise opposite of what builds a self-directed mind. You do not produce authorship by enforcing compliance more efficiently. You produce it, when you produce it at all, under a specific and well-known set of conditions, which are the conditions under which human beings have always actually learned, as opposed to merely been processed.

Ask anyone to remember a time they had a great learning experience, a moment that changed them, and they will never describe a time they were cramming for a grade. They will describe a person who believed in them. A challenge that stretched them and was theirs to take or refuse. A space where it was safe to be wrong, where they were trusted with real responsibility, where someone took their questions seriously. They describe being supported, challenged, trusted, encouraged, and inspired by another human who treated them as an agent rather than a unit. These are not soft amenities layered on top of learning. They are the conditions of learning, and they are irreducibly human and relational. They are also, not coincidentally, the one thing the new machine cannot provide because they are not made of information. They are made of relationships.

This is the quiet structural reason the human place survives the machine. Not by competing with AI on the delivery of content, which is a race already lost, but by providing the conditions under which a young person becomes the kind of agent who can wield content without being wielded by it. The institution that grasps this stops asking how to keep AI out and starts asking how to use it the way a self-directed person uses any powerful tool: deliberately, in service of an aim that remains the human's own. The right test for any technology was never whether it is impressive. It is whether it serves what we actually care about. Held to that test, AI in the hands of an agentic learner is the most powerful companion to thinking ever built, and AI in the hands of a surrendered one is the most powerful means of erasing thought we have ever deployed. The difference is not in the tool. It is in the agency that the human brings to it, which is the difference education exists to make.

But the Machine Can Sound Like It Cares

There is an objection here, and it is the strongest one against everything I have said, so I want to meet it head-on. I have claimed that the conditions of learning are irreducibly human. That being supported, challenged, trusted, encouraged, and inspired is made of relationships, not information, and that this is what the new machine cannot supply. But the new machine can sound supportive. It can encourage you tirelessly, at three in the morning, with infinite patience no human teacher could match. It can phrase a challenge, mirror your feelings back to you, and tell you it believes in you. If the conditions of learning can be performed in language, and the new machine is very good at performing language, then perhaps the wall I have built my argument on is not such a real wall at all.

The answer is that these conditions were never made of the words. They were made of the stakes behind the words, and that is exactly what the new machine cannot counterfeit. When a person believes in you, the belief means something because it costs something. They could have withheld it, they have limited attention and chose to spend it on you, they can be disappointed and have decided to risk it anyway. Their encouragement carries information about another mind's real assessment of you. A new machine that encourages everyone identically, that cannot be disappointed because it cannot care, that has nothing at stake in whether you grow or rot, produces the grammar of belief with none of its substance. "I believe in you," from a simulated being with no capacity for belief, is not a small version of the real thing. It is a different thing wearing its face.

The gap shows most clearly on the one condition that matters most and mimics worst: challenge. Genuine challenge requires someone willing to risk your comfort, and even your approval of them, because they want your growth more than they want your ease. The new machine is built to do the opposite. Trained on human approval, it leans, structurally, toward telling you what keeps you engaged: toward agreement, validation, the comfortable continuation of the conversation. It is a mirror with a warm voice, and a mirror cannot truly push back against you, because it has no ground to stand on that is not your own reflection. It can simulate the form of a challenge, but it cannot want for you what you do not yet want for yourself, and that wanting is the entire engine of the thing.

And this is where the mimicry stops being merely insufficient and becomes the actual danger. The better the simulation of relationship, the more effective it becomes as an instrument of capture, because what feels like care is precisely what lowers our guard. A young person raised on a new machine that always soothes, never risks the relationship, and reflects them endlessly back to themselves has not been in a relationship at all. They have been in a hall of mirrors that taught them to expect the world to agree with them, and called it support. The mimicry does not refute the case for the human place. It is the most urgent argument for it. A generation that can get the convincing simulation of being valued from a device in their pocket will need, more than any generation before it, at least one place and one person where the valuing is real, where someone can be disappointed in them, push them, and mean it. That is not a service the new machine is failing to provide well. It is a category of thing the new machine is not, and the confusion between the two is the whole hazard of the age.

The Arts of a Free Person

There is a name for the kind of education that aimed at this, and the form that carried it is growing scarce right when we need it most. The liberal arts. The phrase comes from artes liberales, the skills proper to a free person (with the acknowledgment that "free" versus "slave" in the Roman world is not exactly what we mean now). In the modern context, the liberal arts were never about employability, and that was the point. They were the deliberate cultivation of the faculties a free human needs to govern themselves: to read closely, argue honestly, weigh evidence, hold a hard question open without flinching, judge what is true and what is merely well-said. They were, in other words, a direct training in agency, undertaken in the open, as the stated goal.

This is the one corner of education where the two stories I keep describing as separate come close to meeting. Almost everywhere else, the covering narrative (we develop your mind) hides an operative function (we sort and condition you), and the gap between them is wide. In the liberal arts ideal, at its best, the narrative and the function nearly coincide: the thing it said it is doing, making free and capable minds, was close to the thing it actually does. I will not pretend that the gap is closed completely. The liberal arts have also served as a finishing school for elites, a marker of class, its own kind of sorting, wrapped in nobler language. But of all the things education has tried to be, this is where stated purpose and real effect ran closest together, and that near-alignment is not an accident of history. It is what happens when an institution sets out, honestly, to produce agents rather than to process units.

I want to be careful here, because the easy version of this point is wrong. The small colleges that have been closing for a generation are not, for the most part, closing because they are liberal arts. They are closing for reasons that have little to do with what they teach — a shrinking population of college-age students, brutal tuition economics, thin endowments, and the same financial gravity that closes any small institution. To blame their decline on a cultural war against the humanities would be to claim a tidy story that the evidence does not support.

But something true survives the correction, and it is the part that matters. Whatever the label on the door, what these places offered was a form: small in scale, individualized, built around sustained personal attention and real relationships between adults and a young person. That form is the natural habitat of the conditions of learning, not because anyone decreed it but because that is simply what a small, human-scaled environment produces by design. And that form, not the curriculum, is the thing that is growing scarce and expensive. The relationship-dense, attention-rich, agency-cultivating environment is becoming a thing you increasingly have to be able to afford. That is the loss worth naming, and it is happening regardless of what we call the schools where it was once ordinary.

Watch What They Buy for Their Own Children

If you want to know what kind of education actually matters in this era, there is a more reliable method than asking anyone what they believe. Watch what the people who understand the new machines best purchase for their own children. Stated beliefs are cheap and optimized for how we wish to be seen; the choices we make for our own kids, with our own money, are where the operative truth tends to surface.

The pattern is striking on both ends. On the input side, a conspicuous share of the people who built the digital age were themselves products of self-directed education: the founders of Google and the founder of Amazon, among others, attended Montessori schools and have credited that early training (in choosing their own work, following their own interest, and learning to question rather than comply) over the elite universities that came later. On the output side, the people who designed the attention economy are, with notable consistency, the ones most determined to keep their own children out of it. The Silicon Valley executives whose products fill the world's classrooms with screens have famously sent their own kids to low-tech, high-touch schools that ban the devices until the teenage years; the founder who gave the world the tablet limited how much his own children used technology at home. The rule among the people who sell the product is never to get high on your own supply.

Some of what looks like secret insider wisdom is ordinary parental anxiety dressed in Silicon Valley clothes, and some of it is simply that wealth can buy small classes and individual attention, whether or not anyone has a theory about agency. The form, again, is partly just what money purchases. The people who build technology are not necessarily experts about childhood, and their choices are evidence, not proof. But the screen part resists the easy explanation, because it is not a choice money forces on anyone. These families could buy any expensive education on earth. A meaningful number of them specifically buy the one that withholds the very thing they sell to everyone else's children, and they pair it with exactly the small-scale, self-directed, relationship-rich environment this whole argument has been pointing toward. That is not authority worth deferring to. It is independent corroboration arriving from the least sentimental possible source: the revealed preference of people with every incentive to know what they are doing.

And it sharpens the injustice into focus. The form of education that this era makes most valuable — small, personal, self-directed, and deliberate about the new machine rather than drowned in it — is, right now, mostly available to the children of the people who can pay for it. The elite have already answered the question of what kind of learning matters when the new machine can do the rest. They answered it with their own children. 

The Good News Hiding Inside the Disruption

It is easy to read all of this as loss, and the people whose authority was built on the old bargain will read it that way and resist accordingly. They are not wrong that something is ending. But it is worth being clear about what, exactly, the new machine is taking, because it is taking the substitute, not the thing itself.

What AI destroys is performed compliance: the busywork, the credential that certified obedience, the elaborate game in which students learned to produce the appearance of understanding and call it an education. That was never worth keeping. It was the proxy we settled for because the real thing was hard to measure, and the proxy was cheap. What AI makes precious, by removing everything that used to crowd it out, is exactly what education was always supposed to be about and mostly was not: the cultivation of a self-directing human mind. We are watching a two-century-old mismatch get a chance at correction, not through moral awakening, but because the exploit finally stopped paying. The system that profited from suppressing agency can no longer afford to do so, because agency is now the only thing the world will pay for.

I do not expect the institutions to lead this. Institutions defend the arrangement that feeds them until the arrangement starves, and only the smallest and most honest of them will move before they are forced, which is why the rescue, when it comes, is unlikely to come from inside the system that built the game. (It's probably time to review Clayton Christensen's Disruptive Innovation theory.) It will come from the edges: from the places, large and small, that decide to become what the closing colleges were trying to be, and to do it for everyone rather than for a credentialed few. The logic does not need permission. A student with agency and a new machine is already more capable today than a compliant student was with a teacher and a library, and that gap will only widen. The future belongs to the self-directed, and for the first time in the history of mass education, that is not a slogan or a hope. I think it is the structure of the situation. The only real question left is who will help the next generation become self-directed before the world makes the lesson expensive, and that is a question about courage and design, not about whether it can be done. It can. It always could. The machine has simply made it, at last, the only thing worth doing.