A Brown University professor recently made headlines with an accusation: half his students, he claims, are cheating with AI. The ensuing debate has fixated on student integrity. I find myself drawn to a question far less often asked — when the tools of an age have changed, who draws the boundary of “cheating,” and in whose interest is it drawn?
Humanity did not arrive at the present empty-handed. Fire, writing, the abacus, the printing press, the calculator, the spreadsheet — each of these, at its birth, was denounced by traditionalists as a dilution of “real ability.” Socrates worried that writing would corrupt memory; today, no one calls consulting a book an act of fraud. To wield tools, to tame them, to co-evolve with them — this is perhaps the deepest endowment of our species. AI is merely the newest link in that long lineage, though it may prove the most powerful.
A university exists, in the end, to serve its students’ futures. And that future is already in plain view: the moment they leave campus and take a seat in any office or at any trading desk, AI will be the most ordinary of companions at their side — as Excel was for the last generation, and the fountain pen for the one before. The question an examination ought to answer, then, is whether a student can command the most powerful instrument of the age — not whether he can perform in a sealed room, offline, in an elaborate pretense that the age does not exist.
I know the objection, and it deserves to be taken seriously. In the foundational disciplines — theoretical physics, pure mathematics — breakthroughs depend on intuition, on mental models that form only through working the derivations by hand and failing at hard problems in person. That kind of understanding cannot be outsourced to any machine. A student who delegates everything to AI can hand in correct answers while building no edifice of his own, and will have nothing to stand on when he reaches problems that have no answer key. I agree with every word of this. But far from refuting my position, it points the way to the real answer: assessment should be neither blanket prohibition nor blanket permission, but a layered arrangement. The foundation must be inspected bare-handed — to certify that what cannot be outsourced has genuinely taken root in the student. Yet inspection cannot be the terminus. Above the foundation, one must eventually be allowed to build with instruments in hand, and to be measured on how far one can go with them. The closer a discipline sits to the applied end of the world, the greater that second measurement should weigh. Nor, for that matter, need AI be banished even from the foundation-laying stage: a tutor available at any hour, endlessly patient, calibrated to each learner, can multiply the efficiency of that bare-handed apprenticeship. The tool can be the finest of coaches in the learning, and need only step aside, judiciously, in the testing. This is precisely why prohibition is the most expensive form of evasion — it cannot be bothered to make these distinctions, and flattens every layer with a single decree.
I do not write about this from the shore. I read Accounting and Finance at King’s College London, where our programme still leans heavily on closed-book, in-person examinations. In two of them — Capital Markets and Asset Management — my results were poor. Let me be honest about what those marks do show: that I failed to adapt to this mode of assessment, and perhaps that there are genuine gaps in my theoretical articulation, in the architecture of my formal knowledge. But I am equally certain of what they do not show: the capabilities I have built in real markets — in research, in live trading, in the management of risk — capabilities on which I have since founded Rocket Capital, a crypto fund of my own, and on which its survival depends every single day. There is an irony here. Those capabilities were acquired in exactly the way a physicist acquires intuition on scratch paper: through the body’s repeated collision with genuine difficulty, tempered by repetition. My scratch paper simply happened to be the overnight tape and decisions made with real money on the line. I have no quarrel with rigor. My quarrel is that the university has been measuring on the wrong training ground.
The problem was never that examinations are worthless. Every ruler measures some true length. The problem is that a discipline whose entire purpose terminates in the real world has taken one limited instrument and treated it as very nearly the sole scale — and then used that reading to define a person’s prospects. What gets misread is not any single grade, but the word “ability” itself.
When I chose to study in Britain, what drew me was the storied openness of its academic culture — a humane environment, so the legend went, that made room for difference and honored more than one path of growth. Having now lived inside it, I can only report a thorough disenchantment. The glory of the tradition endures; the courage to answer the present has yet to arrive.
The Brown professor writes that what alarmed him most was not the cheating itself but his university’s reluctance to confront it. I would say that reluctance is indeed the heart of the matter — except that the party who ought to hesitate is not the student body but the institution. When half of one’s students are reaching for the same instrument, the more urgent question is not why they crossed the line, but why the line is still drawn on an obsolete map.
One more confession is owed before I close: this essay was itself written by AI. The views, the experience, the convictions — every line of thought is mine; the machine’s part was to forge them into prose. I see nothing improper in this. If anything, it is the most honest footnote the whole controversy could ask for. I have never been gifted with words. Under the old order, these thoughts would likely have stayed silent forever, muted by the clumsiness of their vessel — one more instance of ability held hostage by the gatekeeping of rhetoric. Now, for the first time, thinking and writing can divide their labor: mine to get the ideas right, the machine’s to say them clearly. Whoever concludes that this disqualifies the essay is not really dismissing me. He is dismissing everyone who has something worth hearing and lacks the eloquence to say it.
Every genuine advance in the history of education has begun with the admission that the world has changed. The university of the AI age can serve as the last gravedigger of an old order, or as the first namer of new capabilities. Test the foundation bare-handed; test the practice armed. Let the tool coach in the classroom and take its assigned seat in the examination hall, according to discipline and stage. Make the examinations harder, by all means — but harder where reality is hard: in judgment, in synthesis, in everything that remains irreplaceably human when human and machine work as one. That is the rigor this age deserves.
