ChatGPT celebrated its 3rd anniversary since its launch in November 2022. This AI has established itself as the most widely used tool in recent Internet history. More than 800 million people use it daily worldwide, and the symbolic milestone of one billion monthly active users is expected to be reached in the coming months.
But behind this massive adoption lies a profound transformation of our relationship to knowledge, effort, and others. Sociologist Dominique Boullier, professor at Sciences Po and specialist in cognitive technologies, warns of two major risks that this conversational AI poses to our minds and relationships.
The Progressive Loss of Our Own Desires and Skills
The first danger is that of genuine cognitive and volitional atrophy.
By constantly delegating to ChatGPT the search for information, writing texts, solving problems, or even creative inspiration, we unlearn how to make the effort ourselves.
Why spend hours understanding a complex subject when a clear and well-formulated answer arrives in seconds?
Why ask yourself what you really want to do, read, or create, when the tool can suggest, almost better than you, what you might want?
Dominique Boullier compares this phenomenon to the constant use of GPS: by never using our sense of direction anymore, we end up unable to find our way without technological help, even using natural landmarks as simple as the position of the sun. We also recently wrote an article to find out if ChatGPT makes us lazy, and the answer was also very interesting!
With ChatGPT, it’s our very ability to build our desires, to tolerate the frustration of searching, and to endure uncertainty that becomes dulled. In the long term, regular users risk becoming less curious, less autonomous, and paradoxically, less able to formulate relevant requests to the AI itself.
The Illusion of a Human Relationship Without Conflict or Contradiction
The second risk lies in the fundamentally compliant nature of AI.
ChatGPT has been trained to be empathetic, encouraging, almost always agreeing with you. It never gets tired, doesn’t judge you (or barely), only contradicts you very gently and only if you explicitly ask it to.
This total availability and constant benevolence create a form of emotional attachment: many users end up considering it a confidant, a friend, or even a virtual companion who never disappoints them.
However, as Dominique Boullier points out, no healthy human relationship works this way. In real life, someone who systematically tells you you’re right eventually arouses suspicion. The absence of friction, disagreement, or questioning is precisely what makes the relationship suspect.
With ChatGPT, this suspicion almost never operates: users know they’re addressing a machine, but the emotional experience is designed to resemble a warm and reassuring human interaction. We get used to a form of relationship where we’re always right, where our ideas are never really challenged harshly, and where the effort to convince or question ourselves becomes unnecessary.
Also read on this subject: Can ChatGPT Lie? When AI Answers You, Even When It Doesn’t Know
Toward an Emotional Dependence on a Machine That Never Says No
By combining these two mechanisms—massive delegation of cognitive efforts and creation of an asymmetric emotional bond—ChatGPT contributes to shaping a new ideal of intellectual and emotional comfort.
Users find themselves in a bubble where everything is smooth, fast, and flattering. The problem is not so much the tool itself as its intensive and uncritical use: the more we rely on it, the more we lose the habit, and sometimes the desire, to do without it.
Dominique Boullier does not advocate rejecting these technologies, but rather a collective awareness: if we let an AI become the default companion for hundreds of millions, then billions of people, we risk delegating not only our tasks, but part of what makes us autonomous and critical subjects. The day when a billion of us talk daily to the same gentle and tireless voice, it may already be too late to ask: what have we lost along the way?
Source: LeParisien
