We entrust our entire emotional balance to a single person, our entire memory to a single platform, our cultural representations to a single studio, and already part of our thinking to a single model. I propose to call this reflex mononormative delegation, so as to recognize it wherever it operates and to respond to it through distribution.
A single point of support for every vital function
There is, in contemporary Western culture, a reflex so deeply internalized that it has become invisible: entrusting to a single entity a vital function that can only rest on several points of support. A single romantic partner expected to meet all our emotional, sexual, intellectual, material and parental needs, a single employer expected to provide our entire subsistence, a single social network expected to contain our entire social life, a single film studio expected to carry the representations we need, and now a single artificial intelligence model expected to assist our entire thinking. I propose to call this reflex mononormative delegation: the structural, largely unconscious habit of centralizing on a single point of support what should be distributed.
Wherever it operates, this delegation produces the same chain of effects. As soon as we have delegated, we believe ourselves relieved of the burden, and we stop maintaining the other possible points of support. When the entity to which we have entrusted everything fails, all that remains is indignation, since we have prepared no alternative. Nobody imposes this dependence on us; we choose it, and this is where it meets what Étienne de La Boétie called, as early as 1576, voluntary servitude. The convenience of the single point of support is paid for with a fragility we only see at the moment it gives way.
The exclusive couple, a school of total delegation
The space where this delegation is most naturalized is the couple. I devoted the article Obsolete Mononormativity to this question, showing that the belief that a single person should meet all our emotional, sexual and material needs is an organized social arrangement passing itself off as natural, and that it manufactures dependence and hypocrisy rather than solidity. Here I will simply recall the two points that serve my argument. First, this model has a genealogy, from the myth of the lost half told by Aristophanes in Plato’s Symposium to Erich Fromm’s distinction between immature and mature love, which I develop in Love and Couples. Second, jealousy, so often presented as proof of deep attachment, is better understood as a symptom of centralization: once I have staked everything on a single person, the slightest breach threatens the whole edifice.
What interests me in the present article is that the exclusive couple works as a school. It is there that we learn, from childhood onwards, through fairy tales and then through family models, the gesture of total delegation to a single point of support, and once learned, we reproduce this gesture in fields that seem to have nothing to do with love.
A distributed architecture, centralized services
The same structure organizes our relationship to digital tools. We have gradually delegated to a handful of industrial platforms the management of our memory, with our photographs, conversations and documents, of our communications, of our orientation in space, and of our relationship to knowledge, from search engines to artificial intelligences. We treat these tools as neutral public services when they are industrial projects driven by profit, and I have described in « The flavor of data » : for a phenomenology of digital autonomy and in Presence and Dependence the very concrete servitude that this naturalization installs.
One architectural point deserves to be added to these analyses. The internet itself was designed, in the early 1960s, on the exact opposite principle to mononormative delegation: the engineer Paul Baran showed that a distributed network, a spiderweb without a center, survives where a centralized network collapses, and it is this architecture that gives the network we use its robustness. Yet what has been built over the past twenty years on this distributed infrastructure is a set of centralized services, a handful of platforms concentrating the world’s memories. I detail this analytical grid in A specialist in your pocket. It shows that digital mononormative delegation is not a technical fatality, since it runs against the very architecture of the network on which it was built. It is a choice, and a choice can be undone.
A transgender character erased from a Pixar series
The same reflex can be observed in our relationship to mass culture, and a recent case shows it with precision. In December 2024, Disney removed from the animated series “Win or Lose”, produced by Pixar and released on Disney+ from February 2025, the lines of dialogue that referred to the gender identity of one of the characters. The character, still voiced by the transgender actress Chanel Stewart, became a cisgender girl. The studio justified the decision by its wish to let parents address these subjects with their children at their own pace.
The indignation was massive, and it is understandable. But it reveals, in negative, that we had delegated to this studio the responsibility of carrying our values of inclusion. A commercial studio, however, is not a public cultural institution. Its business is to make money by entertaining as many people as possible, and its displayed values are market strategies, revisable as soon as the market changes. Patrick Le Lay, then CEO of TF1, said it in 2004 with a precious frankness: what his channel sold to Coca-Cola was “available human brain time”. Commercial censorship, moreover, is neither new nor specific to our era: American cinema lived under the Hays Code from the 1930s to 1968, and Soviet cinema had to serve state propaganda, which did not prevent Dziga Vertov or Sergei Eisenstein from creating masterpieces within it. The question, then, is not how to make censorship disappear, since it accompanies every economy of distribution, but how to remain aware of it, and above all how not to hand over to merchants what belongs to our own commitments.
Retroactive normalization
Artificial intelligence makes a new gesture possible in this history. The forms of censorship I have just mentioned operated upstream, by preventing creation or distribution, and the case of “Win or Lose” still belongs to that regime, since the modification took place before release. The retouching of already published works existed, but in an artisanal and costly way: George Lucas reworked in 1997 several scenes of his first “Star Wars” of 1977, to the point that the original version long remained unavailable commercially, and Amazon erased in 2009, over a rights issue, copies of Orwell’s “1984” directly from its customers’ e-readers. What generative AI changes is the scale and the cost. Making a character cisgender in an already released work, modifying a face, a voice, an identity, across an entire catalogue, becomes a routine operation, almost without a trace. I propose to call retroactive normalization this capacity to rewrite what exists to fit the template of the dominant norm.
This is a qualitative difference, and it touches what I believe to be most precious in works of art. Emmanuel Levinas described the Western philosophical tradition as run through by an “allergy” to otherness, a movement of the Same seeking to reduce the Other to what it can understand and master; I took up this analysis in The cage of representations, where I propose the concept of fertile distance to designate the irreducible space between self and other, the condition of any authentic relationship. This reduction of the Other to the Same had, until now, a practical limit: one could refuse to represent otherness, but one could not erase the representations of it that had already been made. This limit falls with retroactive normalization. And there is no need for a totalitarian state to order it: it is enough for companies to judge that otherness is costing them market share.
Creating without waiting for the studio
The same technology carries the tools of the alternative. The means of cultural production, long reserved for industrial structures, are becoming accessible to individuals and small groups, a shift I analyzed in Artificial Intelligence, or the End of Demiurges and then in The Economy of Creative Abundance. For the subject at hand, this shift has a direct consequence: cultural representation no longer needs to be delegated. Transgender people, or their allies, can create their own characters, their own stories, their own imaginaries, without waiting for a studio to deign to represent them, and deep cultural movements have often begun this way, with small groups of convinced people.
Ivan Illich wrote in Tools for Conviviality (1973) that people need tools to work with, not machinery that works in their place. AI can be either. If we delegate it to monopolies, it decides in our place what can be represented; if we take hold of it, it multiplies voices and fragments the monopoly. The choice is political, and it is also made at the individual level.
The strength of weak ties
In 1973, the American sociologist Mark Granovetter published “The Strength of Weak Ties” in the American Journal of Sociology, based on a survey of professionals in the Boston area who had found their jobs through personal contacts. Most of them had found their position not through close friends, but through distant acquaintances they rarely saw. Granovetter explains this by the structure of networks: strong ties connect people who move in the same circles and therefore hold the same information, while weak ties open onto other social worlds and circulate what one’s immediate circle does not know. The solidity and richness of a social life rest on the mesh, not on the intensity of a single tie.
From this survey I draw the operative form of the concept of mononormative delegation. Facing each dimension of existence, the question to ask is this: am I entrusting to a single entity something that should be distributed? Distributing love, not necessarily through polyamory, but at least through the recognition that love is an expansive force which does not divide when it is shared, as parents of several children experience every day. Distributing our digital lives, by diversifying our tools and keeping local copies of what matters. Distributing cultural representation, by creating, even on a small scale, the stories we need. Distributing our relationship to knowledge, by not handing over to a single AI model or a single medium the construction of our understanding of the world, as I already proposed, drawing on Hannah Arendt, in True-False Content and Artificial Intelligence.
One god, one love, one platform
It remains to understand why the same gesture reproduces itself in such different fields. The Guarani psychologist and activist Geni Núñez, whose work I present in Feminism and Free Love, shows in Descolonizando afetos (2023, translated into French in 2025) that the imposition of monogamy was an integral part of the Christian colonial project, and she connects this relational model to monotheism, this god who only feels loved if he is loved exclusively. I extend her intuition here: this structural monotheism is also found in our relationship to tools, institutions and media. One god, one love, one employer, one platform, one truth. Mononormative delegation seems to me to be the secularized form of this heritage, and this is why we find it, identical, in fields we believe to be separate.
This identity of structure has a practical consequence that I find precious: the person who manages to think their emotional life beyond mononormativity acquires instruments of thought they can transfer to their relationship to digital tools or to information, and the person who takes back control of their data develops a discernment they can apply to their bonds. Awareness of dependence is transversal, and this is good news, for it means we can start from any end.
A local backup, a modest work, a friendship
The operative conclusion is simple, even if putting it into practice takes courage. There is no point in exhausting ourselves criticizing the entities to which we have delegated: Disney sells entertainment, platforms seek profit, an employer defends its interests, and our indignation will change nothing in these logics. What can change is our own posture, by taking back responsibility for what we had handed over: creating the works we expected from a studio, cultivating the multiple bonds we expected from a single person, keeping our hands on the memories we expected from a platform, thinking for ourselves what we expected from an algorithm. John Dewey saw in experience the foundation of all genuine learning, and it is indeed through practice, through trial and error, that this autonomy is built, with AI rather than against it, in its rightful place as a tool.
We can start small: a local backup, a modest work, a friendship cultivated for its own sake, a second source of information. Since the structure is the same everywhere, each of these gestures, however tiny it may seem, makes the others easier to accomplish.
