Until recently, a familiar scene played out at universities everywhere. A student could not solve a problem. They tried. Failed. Tried again. Asked another student. They argued about it. Neither knew quite enough. Eventually a professor, teaching assistant, or enlightened soul who had understood page 37 came to the rescue.
It was slow, uncomfortable and wonderfully inefficient.
It was called learning.
Artificial intelligence has introduced a small revolution into that ritual. Today, only seconds may separate a question from an answer. The problem is no longer necessarily finding the solution. It is determining whether the person receiving it learned anything along the way.
The Massachusetts Institute of Technology —hardly an institution that can be accused of hostility towards technology— decided to take that question seriously.
Its examination of AI in teaching, learning and research training points towards a remarkably sensible principle: do not declare war on the machines.
Defend the human being.
Because there is an enormous difference between using a tool to think better and using it to avoid thinking.
The first expands human ability. The second may gradually weaken it.
Calculators save us from mechanically repeating certain calculations. Search engines find in seconds what once required hours in a library. Nobody sensible proposes returning to the abacus or switching off the Internet to save civilisation.
Artificial intelligence, however, introduces something different.
It does not merely calculate or search.
It argues. Summarises. Connects. Explains. Writes. Suggests. Corrects.
Above all, it answers.
And there lies the delicate frontier. A student can use that answer to advance their reasoning — or surrender to the machine precisely the reasoning they were supposed to exercise.
It is the difference between using a walking stick and asking somebody else to walk for us.
That brings us to a particularly troubling idea: cognitive surrender.
This does not mean using artificial intelligence. It means giving up intellectual effort too quickly and accepting an answer before developing enough judgement to evaluate it.
Making mistakes, after all, has acquired a terrible reputation.
Yet much of human intelligence was built through mistakes.
A student who solves a problem incorrectly and discovers why may learn considerably more than one who instantly receives the perfect solution. Frustration, doubt, trial and error —even that inelegant moment spent staring at the ceiling waiting for an idea— are part of learning.
The machine can spare us that discomfort.
And therein lies the danger.
The answer cannot simply be to turn every classroom into a technological checkpoint, hunting suspicious prose and placing blind faith in automated AI detectors.
There would be a delicious irony in combating excessive artificial intelligence by deploying another artificial intelligence to determine whether students used artificial intelligence.
All we would need then is a third machine to supervise the first two.
The harder and more serious path is to reconsider education itself: what we teach, how we teach it and, especially, what it means to prove that somebody has learned.
If a machine can produce an impeccable essay, perhaps the submitted essay is no longer sufficient.
Talk to the student. Ask them to explain. Defend. Connect. Show how they reached the conclusion. Bring back projects, practical experience, discussion, community and shared work.
When it becomes difficult to establish who wrote the answer, perhaps we should concentrate again on who understands the question.
And at that point the issue leaves the university campus.
Cognitive surrender does not threaten only the twenty-year-old with an exam on Friday.
It threatens all of us.
The journalist who asks for the conclusion before reading the documents. The lawyer who accepts a precedent without checking it. The official who reads the summary and stops opening the file. The citizen who requests a ready-made opinion to avoid the inconvenience of forming one.
And, naturally, those of us who work with artificial intelligence every day.
The question should never merely be how much the machine can do.
Every month it will be able to do more.
The truly uncomfortable question is what we should continue doing ourselves even when the machine can do it better, faster and without asking for coffee.
Thinking may technically become optional.
But a society that begins to regard it as unnecessary will have confused progress with surrender.
Artificial intelligence can be an extraordinary travelling companion. It can broaden knowledge, reveal hidden connections and place once-specialised capabilities in the hands of millions.
But the travelling companion should not end up occupying our seat.
The more intelligent our machines become, the more important it will be to teach human beings not to surrender to them the very thing that makes them human.
We do not need students capable of competing with artificial intelligence.
They would probably lose.
We need people capable of using it, challenging it, doubting it, correcting it, benefiting from it and, when necessary, closing the screen and saying:
Now it is my turn to think.
🖋️ © El Testigo del Tiempo | 2026 — All Rights Reserved© 2026 SalaStampa.eu, world press service – All Rights Reserved – Guzzo Photos & Graphic Publications – Registro Editori e Stampatori n. 1441 Turin, Italy
Sources: MIT — Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training; MIT Office of the President, 25/08/2026