Some time ago, I was exchanging messages with my partner about something personal and particularly sensitive. In the middle of the conversation, a visual detail made me stop reading: a sudden overabundance of em dashes (—) in her sentences.
For anyone who works daily in the engine room of artificial intelligence, this is not a simple typographic choice. It is a fingerprint. It is the unmistakable syntactic signature of a generative large language model (LLM).
I immediately asked her: "Did you use ChatGPT to reply to me?"
She admitted that she had, assuring me it was "just to correct her spelling mistakes".
The explanation was reassuring, but the poison of doubt had just been injected. What if the reason was more insidious? What if, faced with the emotional friction of a sensitive subject, she had chosen, out of comfort or fatigue, to delegate the formulation of her thoughts to the machine? In that intimate exchange, it had suddenly become impossible for me to distinguish the real from the fabricated: where did her true intention end, and where did the algorithm's smooth, polished, synthetic response begin?
This vertigo is not a science-fiction scenario. It is the new world we have woken up in. Today, anyone can plug a tool like OpenClaw into their WhatsApp and let a probabilistic model manage their social interactions.
As a Senior Applied Scientist building these architectures every day, that intimate moment hit me like an electric shock. I realised that the urgency was no longer purely technical — it was anthropological. I felt a visceral need to lift my head from the code and plunge into the philosophy of technology.
I reread Jacques Ellul, who had already warned us that a tool always ends up imposing its own logic on whoever uses it. Even more disturbing, faced with this massive and silent acceptance of algorithmic assistance, it was the voice of Étienne de La Boétie that resonated. In his Discourse on voluntary servitude, he dissected the mechanics of political tyranny. The analogy with our era is chilling: the tyrant has only the power we cede to him for the sake of convenience. Today, tyranny no longer wears the face of a despot — it has the smoothness of a frictionless interface that spares us the effort of thinking.
We are becoming, in many respects, the embodiment of Nietzsche's prophesied "last man": a tired being, incapable of tolerating the slightest friction, who invents ever more sophisticated tools to secure a numbing comfort. The tragedy is that it is precisely these "last men" who are today designing the algorithms that govern our world.
It is from this vertigo that my essay was born: The laziness of thinking: the great intellectual abdication in the face of AI (forthcoming from FYP Éditions).
I did not write this book to sound a hollow technophobic alarm. AI is a prodigious tool. But it is vital that we strip it of its aura of "black magic". We must stop treating these models as infallible oracles capable of delivering truth on demand. They are statistical architectures, designed to predict the most probable word — not to understand the weight of our silences or the pain of our choices.
As long as we regard artificial intelligence as a magical creature, we will be condemned to be ruled by it. But if we are willing to lift the hood, to understand its intimate flaws and its biases of complacency, then we can take back control.
In this book, I explore how we have come to delegate our free will — but above all, I propose to forge the intellectual weapons to reclaim it. Because faced with the machine, the only true line of defence remains our courage to think for ourselves.