Étienne Klein
“AI doesn’t think — it forces us to think differently.”
With a blend of vigilance and reasoned optimism, physicist and philosopher Étienne Klein urges us to move beyond technological fascination and adopt a more active stance — one of lucid companionship between human intelligence and machine power.
Q: Artificial intelligence is now present in every field. Should we see it as a technological turning point — or a civilizational one?
Étienne Klein: Both, undoubtedly. AI isn’t just another tool — it’s changing how we produce, share, and validate knowledge. It’s an anthropological shift: for the first time, machines are participating in the creation of meaning. But our brains are still wired for old habits — we prefer confirmation over contradiction. In a digital world where every opinion coexists, this fuels what I call the “power of the false”: falsehoods, being more spectacular than truth, spread ten times faster. So the real challenge isn’t just technical — it’s a revolution in critical thinking.
Q: In a world where everything is accelerating, is there still room for reflection?
E.K.: Yes — but only if we make space for it. I see it with my students, and anyone can observe it: those who take the time to think before asking an AI a question gain real intellectual value. Others outsource everything to the machine and let their minds atrophy. AI shouldn’t be a crutch — it should be an extension of our thinking. Otherwise, it impoverishes our relationship with knowledge.
There’s a uniquely human joy in understanding — a deep satisfaction that comes from discovering something on your own, a joy that leaves a lasting imprint on memory. That’s the joy we must protect. It’s part of what makes us human.
Q: Every major technological revolution reshapes our relationship with the world. How is AI forcing us to reinvent ourselves?
E.K.: History shows that technological breakthroughs don’t kill human creativity — they shift it. I often compare it to the invention of photography: when the camera arrived, painters feared for their craft. But that fear sparked the birth of abstraction — a leap into what machines couldn’t replicate. AI presents a similar challenge today: it’s up to us to invent what machines still can’t imagine.
Q: But this human–machine alliance raises a key issue: trust. Can we really “believe” an AI?
E.K.: The word “intelligence” in artificial intelligence is misleading. In English, intelligence often means information processing, not reasoning. Machines process data — they don’t understand. When an AI gives you an answer, it doesn’t explain how it got there. That puts our critical thinking on hold: we either believe it or we don’t, with no way to tell the difference
This is a serious issue, especially when decisions affect lives or society. At the CEA, for example, we monitor tsunami alerts in the Mediterranean. An AI might flag an abnormal wave — but based on what? No one really knows. A false alarm causes panic; a missed one could be catastrophic. Blind trust isn’t an option. We need geophysicists to interpret, argue, and decide. AI isn’t an oracle — it’s a powerful but fragile tool. Meaning and decisions must remain in human hands.
Q: You’ve spoken about a “fatigue of intelligence” among young engineers. What do you mean by that?
E.K.: In France, we’re seeing many brilliant young people — especially in physics and math — turning away from engineering. They’re heading into finance or consulting instead. They feel overwhelmed by technical complexity. The more powerful technology becomes, the less we understand how it works. This opacity breeds a sense of helplessness: we admire the machine, but we no longer master it.
It’s a kind of Promethean shame — pride in having built powerful tools, mixed with fear that we can no longer comprehend them. In a world of black-box systems, our first priority should be to relearn how to explain.
Q: How can we bring back a culture of understanding — in business and in everyday life?
E.K.: Through education and experience. At Centrale, we’ve launched writing workshops with no digital tools. Students rediscover the joy of formulating ideas, of thinking together. It’s not nostalgia — it’s a way to re-anchor human intelligence in the act of thinking.
The same applies in business: we must learn to understand what we use. Teaching AI isn’t just about tools — it’s about method. It’s about learning to question results, to doubt, to argue. These critical skills will define the difference between passive users and augmented professionals. We need to train, experiment, debate — that’s how we avoid becoming mere consumers of artificial intelligence.
Q: You’ve often spoken of your admiration for Albert Einstein. What do you think he would say about AI?
E.K.: I don’t think Einstein would feel threatened by AI — and he’d be right (as usual). When he published his theory of general relativity in 1915, physicists had far fewer data than we do today. Yet he found the right equations.
Now imagine a world where we had all today’s data, but Einstein’s theory had never been discovered. Could the best algorithms, fed with all that data, somehow induce the right concepts and reconstruct his equations? The answer is no.
Q: If you could leave one open question for our readers — business leaders and IT professionals — what would it be?
E.K.: I’d ask: How far are we willing to delegate our intelligence?
True modernity isn’t about handing over decisions to algorithms. It’s about understanding how they reshape our thinking. AI isn’t here to replace us — it’s here to force us to rediscover what it means to think. And that, paradoxically, may be its greatest gift: bringing us back to what truly matters — our capacity for discernment, creativity, and the joy of thinking that nothing can ever automate.
[LEARN MORE ABOUT ETIENNE KLEIN]
Étienne Klein is a physicist and philosopher of science, and a member of the French Academy of Technologies. He heads the Research Laboratory for the Sciences of Matter at the French Alternative Energies and Atomic Energy Commission (CEA), where his work focuses on the philosophy of physics and applied ethics. He teaches philosophy of science at CentraleSupélec and hosts the weekly radio show La conversation scientifique on France Culture.
His recent books include L’éternité béante (Futuropolis, 2024) and Transports physiques (Gallimard, 2025).