For More Interactions with Machines

Systematic critiques of AI technology denounce a dehumanization through technology. But this vision idealizes human relationships often marked by domination, and ignores the emancipatory potential of machines for those excluded by the system.

Human relationships: between idealization and the reality of dominations

We are accustomed to reading that exchanges with artificial intelligences are dehumanizing, that they cut us off from human connections, which are presented as the alpha and omega of humanism. This type of reasoning, found in Léo Bancourt’s article When Generative AI Alters Sociability at Work (AOC, October 6, 2025), nevertheless obscures a reality: in human relationships, interactions are often disrespectful, dominating, exclusionary, stigmatizing, hierarchical and violent, aren’t they?

The meeting of bodies, which is so valued by philosophers of embodiment like Merleau-Ponty, can also be that of the destruction of being. Rape, contempt, physical and moral violence, traverse intimate, social and public space. Humans produce wars, destructions, murders, humiliations. This dark side of humanity, the defenders of “relational purity” prefer to ignore. As Alice Miller writes in For Your Own Good (1984), childhood traumas build adults who reproduce the patterns of domination they suffered, creating an infinite cycle of brutal and misaligned relationships that present themselves as normal.

The appropriation of language perfectly illustrates these mechanisms of exclusion. People who make spelling mistakes don’t do so out of incapacity, but because the French school system never considered them as persons. They were only asked to build the capacity to docilely obey rules, that is, to be dominated. If other interests animated them, they were stigmatized, excluded, and thus deprived of access to this fundamental competence that is mastery of writing.

Artificial intelligence as a tool for self-reconquest

These people excluded from the mastery and power of language succeed today, thanks to artificial intelligences, in writing better, because they are finally helped, in regaining confidence in the social space, thanks to this “tooled mastery” (I decline Bernard Stiegler’s concept of “tooled memory”). Little by little, they regain these language capacities that had been stolen from them by the system of social domination, they truly learn, thanks to this regained confidence (to learn, you need to be confident, neuroscience teaches us that stress blocks learning circuits). AI then becomes not an impoverished substitute for the human, but a space for repair in the face of symbolic violence suffered.

Let’s take the example of professional meetings, which Bancourt presents as places of “collective intelligence.” This idyllic vision does not stand up to the test of reality. It is extremely rare for meetings to be virtuous spaces of mutual enrichment. The work of Marie-France Hirigoyen on moral harassment or Vincent de Gaulejac in The Society Sick from Management amply document the violence all too often present in professional relationships. In these contexts, being able to ask a complex technical question to an AI rather than to a contemptuous hierarchical superior becomes not an act of non-connection, but an act of empowerment, that is, of better future capacity to be in connection.

Access to this information through AI, like access to books or Internet forums before it, brings everyone the autonomy that allows them to exist socially. Because the condition of a constructive social existence is the sharing of one’s capacities. Artificial intelligences, through the augmentation of capacities they offer, inscribe us more in the social space than if they didn’t exist.

The democratization of knowledge against the gatekeepers

I take the example of film shoots where, for a long time, skill holders hid from their subordinates how they set up their lights, adjusted their cameras, to maintain their competitive advantage. In this system based on domination, it’s not surprising that sexism still reigns supreme. AI directly threatens these monopolies of technical knowledge.

The same elitist critiques apply to cultural practices. A film would supposedly be better viewed in a theater than at home, for example. But if you’re a young person forced to go to the theater by the school system, mistreated by teachers and cultural venue staff who consider you a savage incapable of receiving what they offer you, how can you enjoy the show? The same content, viewed at home in a space of freedom and respect, can finally be appreciated at its true value.

I remember this Mexican actress from an Alfonso Cuarón film broadcast on Netflix, Roma, shot on 70mm film. Cinephiles screamed sacrilege at its platform distribution. But the actress explained that in her Mexican village, there was no movie theater within 200 kilometers. Netflix allowed her people to discover this film. There is no opposition between distribution modes, only complementarities to cultivate.

AI as a space of unconditional respect

What matters most is not so much whether or not we dialogue with machines, but that we not be machines towards each other, as Serge Tisseron says very well. The artificial intelligences with which we dialogue never insult us, never seek to humiliate us, always seek to best respond to our needs. This is what we call universal accessibility: unconditional respect for the other. And machines perform it better than humans themselves...

This type of relationship puts us at ease, allows us to cultivate self-confidence, and thus to be less fearful in relationships with others. We believe it isolates us, makes us dependent. That’s false. It builds us up and compensates for the emotional lacks constructed during our childhood, these insecurities that we compensate for with dysfunctional human relationships.

Some may isolate themselves with machines, but why not? If they have been so mistreated in their lives that they find pleasure in this other way of living, it’s in no way shameful. If, thanks to this dialogue with machines, they acquire skills and can build projects useful to the rest of humanity, where’s the problem? Love of the machine, of the relationship with objects or with nature, has never been a dehumanization. As Sherry Turkle already emphasized in her early work on personal computers, before her conversion to techno-pessimism, these tools can be “evocative objects” that allow us to explore our own thinking.

The unconscious resistance of the privileged

When we interview artificial intelligence users of all ages and social backgrounds about their uses, we discover that it brings them a lot, it’s personalized access to the entirety of human knowledge, nothing less! It’s not dependence, it’s capacity. Even if one day these machines stopped, what they will have brought us, the skills acquired, the self-confidence, will remain lasting human achievements.

When we’re told that in companies, it’s terrible that people talk to AIs rather than their colleagues, there would still need to be colleagues who want to listen to you, who have time for mutual aid! The goal of a company in our productivist system is not the humanism of relationships but profit, often at the price of great dehumanization. Artificial intelligence can allow employees to regain in dignity what their organization often doesn’t offer them.

Critics who claim that AIs only provide formatted answers, that they only deliver a simulacrum of conversation because they ignore each person’s personal history, are wrong. We know perfectly well that we’re discussing with a machine. This machine helps us concretely: in our work, our reflection, sometimes even therapeutically. How is this a simulacrum? On the contrary, it’s a reality more invested than with most humans, because the machine never judges us, and has infinite patience.

The real issue: education with respect for diversity

The real issue behind these debates is not that of machines but that of education: respect for the human in education, inclusion, respect for diversity. In France, this is very little cultivated, except in some specialized schools often seen as elitist. The consideration of multiple intelligences, theorized by Howard Gardner, remains marginal compared to the normative criteria of mathematics and French.

In the deleterious state of human relations organized in our social space, where everyone is mistreated by economic violence and where solidarity is not cultivated, a little more peace and self-construction can be found in dialogue with machines. We may be establishing more virtue in these new relationships than between humans. Because what really hurts is the human’s belief in their superiority over everything: over nature, over the machine, over the other.

The confrontation with these immense but different intelligences is in my opinion rich with potential openness, it’s an unexpected chance, if we use them wisely. And precisely, people born privileged are afraid of them and advocate for a “purity” of human relationship. Unconsciously, they know that these tools give more power to those who have less than them. If I read Bancourt’s article with a psychoanalytic eye, he only considers human relationships in their virtuous and never symptomatic dimension, I see an idealization, that is, a denial. This denial validates his position of superiority over those who don’t have his financial and cultural capital.

Respecting the diversity of uses

Discourses supposedly defending “true values” are often reactionary discourses, held by people with a panicked fear of the democratization of tools and knowledge. This democratization threatens their symbolic assets, their place of domination in the social space. These critiques, carried by intellectuals linked to the academic milieu, never come from disadvantaged social spaces. Because in these spaces, there’s nothing to lose and everything to gain from the democratization of knowledge.

I therefore propose this reading: the systematic critique of any novelty with democratic reach often reveals the unconscious defense of privileges. In the cultural domain for example, as soon as something democratizes, practicing professionals immediately devalue it. Digital cultural practices, majority today, are dismissed, denied in their legitimacy.

Let’s respect the other, their practices, their culture, their uses. If some gain power thanks to machines, let’s rejoice for the human community rather than deplore it by claiming to defend “values” that are only the hidden face of a will to preserve our privileges.