Published in Quadrant (Sydney), 25 June 2026
A sense of dread has settled over schools and universities. A machine can now read in an afternoon what a scholar could not finish in a lifetime. It writes papers, works problem sets and can pass many exams in seconds. The skills a degree certifies can be imitated by anyone with an AI subscription, at little cost. Teachers and academics have begun to ask, half under their breath, whether their work is about to become redundant.
The dread is born of confusion about what schools and universities are for. It takes the purpose of education to be the paper and the exam, the thing to be marked. But the paper was never the point. The point is the students, and what writing the paper made of them.
Yet the debate inside the institutions runs on contents and conduct. Which skills will still earn a wage once the machine has taken over the rest? How do we stop students passing its essays off as their own? Should we teach prompt engineering, whatever that will mean by next year?
These questions treat education as useful contents poured into a mind and processed into an outcome, with the machine as a rival supplier. To see past that picture, it helps to have a word for what it leaves out. German uses two words to make a distinction that English tends to blur, and that goes to the heart of educational purpose.
The first word is Bildung, and no English equivalent has ever quite stuck. I grew up with the word and the concept behind it. In German it is simply in the air of a school, the tacit reason the place exists. It took leaving Germany to notice that English had never had a word for it, nor quite the ambition behind it.
So, what is this Bildung? Bildung is not qualification, nor quite self-cultivation, which sounds like a hobby anyway. No, it means the forming of a whole human being, the development of a person’s own powers into something complete and uniquely his own.
The second word sounds similar but means something different. Ausbildung, training, is the fitting of a person to a task. It is necessary, even admirable. Still, in English-speaking countries, it has been mistaken for the whole of education.
You can see the confusion playing out in anxiety over students using AI to do their assessments for them. The handwritten, invigilated exam is coming back in some Australian and British universities, to make sure it was the student and not the machine who performed the task.
We have decided the problem is one of policing. What we are checking is whether the student, not the machine, can execute what the exam asks, which is to say whether the Ausbildung worked.
But that is only what policing can see. It catches whether the training has done its work, never whether a person was formed. Set against the broader ideal of Bildung, the machine is not a threat at all but a test of it.
The man who put Bildung at the centre of a modern education system did it in sixteen months, while his country lay in ruins. Wilhelm von Humboldt was born in 1767 into a recently ennobled Prussian family. He moved in the same circles as Goethe and Schiller, poets of a German flowering of letters and liberal thought. This is the Germany the world forgets when it remembers only the Prussian parade ground.
Humboldt was a polymath. He read law and classical philology at Göttingen, translated Aeschylus and made pioneering studies of Basque. He gave years to the Kawi language of Java, sure that how a people speaks shapes how it sees. He served as Prussian envoy in Rome for six years.
Then, in 1809, Berlin called him home to a broken Prussia, half-occupied by Napoleon. He took charge of education in the interior ministry for barely sixteen months, but it was time enough to redesign Prussian schooling from the elementary classroom up and to oversee the founding of the Berlin university that still carries his name.
Humboldt was a living argument for Bildung. Here was a mind at home in law, the classics, many languages and the politics of a continent. He was at ease in each and captive to none. He was the embodiment of the gebildeter Mensch, a well-rounded and well-formed human being.
At the centre of Humboldt’s philosophy stands the person who, “without any aim directed at anything particular”, wants only to raise the powers of his own nature. The highest task of our existence, he wrote around 1793, is to give the concept of humanity in our own person the greatest possible content. It happens only through the freest interplay, the Wechselwirkung, between the self and the world.
Knowledge mattered to him, not as a stock to be banked but as what the self feeds on and grows from. Learning that never fed back was, to him, just so much dead weight.
He had put it even more plainly as a young man, in the book later translated into English as The Limits of State Action, whose argument was that the state should leave men alone to form themselves. It was classical liberalism before the phrase existed, and Prussian censorship kept it from full publication, which rather proved its point. Only fragments appeared in his lifetime, and the whole book was not published until 1851, sixteen years after Humboldt’s death.
One of those fragments was the chapter Schiller ran in his journal Neue Thalia in 1792, where Humboldt set down the line. The true end of man, he wrote, is “the highest and most harmonious development of his powers to a complete and consistent whole”.
It rests on two things. The first is freedom. The second he thought no less vital, a real and demanding engagement with what he called “a variety of situations”, a world that pushes back. For “even the most free and self-reliant of men is hindered in his development, when set in a monotonous situation”.
Then comes a sentence that, read today, seems to describe a screen. Whatever “does not spring from a man’s free choice or is only the result of instruction and guidance”, he wrote, “does not enter into his very being, but still remains alien to his true nature; he does not perform it with truly human energies, but merely with mechanical exactness”.
Of course, Humboldt meant instruction and authority where the answer is handed down rather than worked out. Yet nothing describes the machine better. It also delivers something from outside the self. With a machine, the answer to almost any question comes back fluent and apparently complete, but it will remain foreign to you unless you take the time, not merely to understand it, but to assimilate its meaning into your worldview.
It would be easy to file all this away as a German peculiarity, the sort of high-minded idea that does not travel. But it did travel. Humboldt’s conviction took root in England in one of its central liberal thinkers.
John Stuart Mill set a line from Humboldt’s book at the front of On Liberty in 1859: “the absolute and essential importance of human development in its richest diversity”. He built his chapter on individuality on it. A man must form himself rather than be moulded, and the forming is the point, not a polish on the useful business of living. That is pure Humboldt, if you like.
Mill was not alone amongst English thinkers in this. John Henry Newman, in The Idea of a University, made the cultivation of the intellect an end in itself, and Matthew Arnold, reading the Germans, set against English philistinism the thing he called “culture”. So, Humboldt’s idea certainly crossed the Channel, but the word Bildung and its inward reach did not.
England adopted the result of Bildung, but not the labour that produced it. It prized the finished article, the gentleman of character, conduct and taste, but quietly set aside the slow inward forming that had made him. The German question was what an education does to a person. The English one was what it leaves him able to show for it. The English have always been more pragmatic than those constantly brooding, permanently pondering Germans.
But even in Germany, Humboldt’s ideal came with its own disease. Bildung could turn into pure inwardness, a man perfecting his soul and his library while leaving the public square to others. Later, more statist and inward versions of German formation drifted far from Humboldt’s own liberal suspicion of power—far enough that a nation of Denker and Dichter, thinkers and poets, committed some of the worst atrocities the world has ever seen.
That can look like a refutation of everything Humboldt stood for, but it is better read as a warning. Inwardness of that kind has cut itself off from the second of his two conditions, the variety of situations, the world that pushes back. Bildung severed from that world does not guard against barbarism. Perversely, it can even dress it up as respectability.
Humboldt the linguist, the scholar of Basque and Kawi, gave his last years to language. Into it he carried the same conviction that ran through all his thought, that a person, like a people, is not made from without but forms itself from within. What he found there bears on our own machine age more directly than anything else he wrote. Language, he held, is not a finished product but an activity.
He called it energeia and not ergon, the “doing” rather than “the thing done”. Language is the mind’s endless forming of its world, not only a mirror held up to it. To think, then, is to build a thought rather than to receive one.
A mind that has done the work brings a world of its own to what the machine returns and can judge it. A mind that has not simply takes the machine’s word for it.
Humboldt built that conviction into his schools policy. General formation came first, and it was the same for every child, the future tradesman no less than the future minister. A man is “only a good craftsman, merchant, soldier and man of business”, he wrote, “if he is, in himself and without regard to his particular calling, a good, decent and, by the standards of his station, enlightened human being and citizen”. Training for a trade was to follow that formation, never to be folded into it.
Something of that order survives in German schooling. It is strongest in the Gymnasium, the academic upper tier. But even the schools that lead to a trade still hold, in principle, that the person comes before the calling, so that Bildung is not meant to collapse into mere Ausbildung.
Many contemporary education systems now do the opposite. They put training first and treat formation as an optional extra, and the machine pushes that choice to its limit. Humboldt would say we have it backwards. We keep asking what the machine will do for us, or to us. But the real question is what we bring to it, and what we do with its products.
To a formed mind, the machine is an amplifier. It can spar like a faculty of colleagues, testing a claim and pushing back. It can fetch and sort like a tireless research assistant. It can make a tangled paragraph plain to a stranger, as a good subeditor does.
But every amplifier needs a signal, and each helper only multiplies what is already there. To a mediocre mind it adds little, and to the pretender only more fluent pretence, which ironically he will be the last to notice. The less a person brings, the less any help gives back, whether from a human or a machine. What separates one set of results from another was never the tool but the formed person using it, the gebildeter Mensch who knows what to ask and what to keep.
The panic over cheating is characterised by the same confusion. Students who let the machine write their essays without learning to think for themselves have cheated no one but themselves.
The machine does not raise or lower us all alike. It does most for the person who needs it least, and least for the person who might seem to need it most. The early studies seem to cut the other way, but only at first glance. When the machine is built to teach rather than to answer, or set alongside a human tutor, weaker pupils can gain much from it. That is Ausbildung doing its proper work.
But where the machine simply hands over the answer, the gain is a loan against the future. In one large school maths trial, pupils did better while they had an unguarded tool but worse once it was taken away, falling below classmates who had never used it. The training gap may narrow while the gap that matters more, the one in formation, quietly widens. A pupil can score higher and think less.
This is why artificial intelligence makes Bildung more necessary, not less. There is nothing sentimental about this. A powerful instrument does more harm than good to a mind never formed to use it wisely. Humboldt’s “complete and consistent whole” is no museum piece. It is what now decides whether you run the machine or it runs you.
But there is a catch. The same machine that makes formation more necessary than ever is also the best tool ever built for preventing it. Humboldt’s formation needed freedom and a variety of situations. It also needed something his age took for granted—the hard, solitary work of thinking a problem through.
A mind grows by exposure to resistance, the way a muscle does. The machine is in the business of removing that resistance.
As an economist, I am usually on the side of whatever removes friction, of the efficiencies that let us do more with less. But formation of the mind is the one exception I know of, the place where friction is the point and the struggle is how we come by Bildung.
Yet with AI, a summary of anything is always just one prompt away. With AI, arguments arrive already assembled.
This is a debate much older than one might assume. None of it would surprise Socrates. Yes, Socrates who made the same charge against writing in his time. In Plato’s Phaedrus, Socrates recounts a myth in which an Egyptian king criticises the god Theuth’s invention of writing, arguing that it would breed forgetfulness and give men the appearance of wisdom rather than the reality. From then on, the alphabet, the printed book, the internet, each was going to rot the faculty it was meant to serve. Yet the mind came through every time. So, the fair question is why this time should be any different with AI.
The answer is that the older tools made the intellectual work easier, where this new one offers to spare us the work altogether. Writing weakened memory, sure, but composing a thought still demanded the thought, and print put answers within reach without sparing the reader the reaching. But the machine can take both away and hand us back an answer the mind never had to build. That is a line the many technologies since Socrates have never crossed.
The monotonous situation Humboldt feared therefore turns out not to be a dull life. No, it is a frictionless one, an endless supply of answers that asks nothing of those who take them.
The people who can least afford the loss are the ones who suffer it. A student who knows how to work uses the machine to do more, while one who does not uses it to dodge the effort. The gap between them widens, even though schools exist to close it.
Once again, this is an old story. None of the earlier tools—writing, printing, the digital computer—reached everyone alike. Each did most for those whose minds were already well formed.
That widening gap is not even the worst of it. The real worry is the children growing up with these machines from the very start. Those of us who are older had already built our minds before AI arrived. We have something of our own for it to work with. But children born today will begin from scratch. The slow, difficult work that would once have formed them is the very thing the machine offers to take off their hands, at exactly the age when it matters most.
I do not know how that will end. No one does. But I think the question for today’s children is no longer which of them will get ahead. It is whether their minds get formed at all.
The answer to all this is not to keep the machine out. To a formed mind it is an extraordinary tool. The answer is to understand that friction has a purpose, and to insist that certain work must still be done the hard way, and alone. The handwritten, invigilated exam I earlier called “policing” is not merely policing. To someone who wants only training, it is a checkpoint. But it is also one of the last things that make a student sit and think without help. That is what forms a mind—but students should not only do it when sitting exams.
The fear that AI will make schools and universities superfluous is thus misplaced. The real danger lies in us having forgotten what education is for. If we let children use AI to do their thinking for them and call the result an education, we will raise a generation that can summon answers but not know what to make of them.
The usual response to technology is to train the young for whatever work the machine cannot do yet. When it comes to AI, that would be the deepest mistake of all. It would be a bet placed on a race we cannot win.
Humboldt would have done the opposite. He would insist that we must form the whole person first, and only then build specific skills and the judgment to use them. A person formed that way can pick up any tool, AI included, and stay in charge of it. Humboldt’s idea has aged far better than he might have guessed.
In its own way, the machine has proved Humboldt right. Education was never about what you could get out of a person. On the contrary, it was about forming the person themselves.
We are about to learn the hard way what it costs to have forgotten that, and that there is a world of difference between Bildung and Ausbildung.
