AIthropology Lab
ES EN

Anthropology and digital culture

Fabricated persons

Synthetic figures are not false persons, but fabricated persons. An anthropological reading of how an image comes to act as a person, a moral subject and a bearer of value, through the cases of Lil Miquela, Shudu and Lu do Magalu.

· Martín González Senosiain
Editorial image for the essay «Fabricated persons», on synthetic figures, agency, value and authenticity.

In December 2019, a bodiless image said that it had been sexually assaulted. Lil Miquela —a figure followed en masse on social media, a computer-generated face, a biography written by a company— posted a video in which she recounted, in the first person, an experience of harassment during a shared ride. There is no body here that can be touched, no memory that has not been drafted by someone, no “it happened to me” that points back to a me prior to the script. And yet the video says I.

The dominant reaction was not the one we reserve for a character. It was not, save for exceptions, the response we give to a cartoon, an advertising mascot or a poor piece of digital animation. People grew uneasy, indignant, offended. That register —moral offence— is not the one we keep for fiction recognised as fiction. It is the one we keep for someone who has done something. The scandal was not aesthetic or technical, but moral. And a moral scandal is possible only where a subject is presumed, someone who can be held to account.

What is decisive lies in how the criticism was framed. Almost no one merely said “this is fake” —something already known in advance. What was said, over and over, was something else. A fiction was appropriating real stories, it had invented an assault in order to seem closer, that place was not hers to occupy. The singer Kehlani reproached her directly, accusing her of toying with experiences and traumas that belong to real people. It is worth pausing on the grammar of that accusation, because it says more than its content. One can appropriate only what does not belong to one, and only someone occupying a position from which appropriation is even conceivable can appropriate wrongfully. To reproach Miquela for having appropriated a human experience is, without noticing it, to grant her the standing of someone who might have had —or not had— a right to it. The accusation treats the image as a person in the very gesture by which it condemns her.

This is where this essay begins. Not in the question of whether Miquela “is real” —an exhausted question, and a badly posed one at that—, nor in what exactly happened that December, which matters little. The question is another, and it belongs squarely to anthropology. What kind of person does an image have to be for it to be possible to reproach it morally for having faked a human experience, to give it affection and, at the same time, to buy advertising space from it? Put another way, how is a person fabricated who is capable of receiving, all at once, moral reproach, emotional attachment and commercial value?

Because that is what synthetic figures are, and the word matters. Not false persons, but fabricated persons. The distinction is not a rhetorical game. “False” takes for granted that there exists, somewhere, the true person this one imitates and betrays; it measures the phenomenon by what it lacks. “Fabricated” asks the opposite. It asks after the operations —image, narrative, interaction, platform, market— that make something with no biography of its own come to function socially as if it had one. The marketing gaze asks whether the device works. The anthropological gaze asks what has had to be fabricated for it to work, and it compels us to read these profiles not as campaign shop windows but as social scenes, places where, before our eyes and with our participation, person is produced, bond is produced and value is produced. That is the factory this essay sets out to look at closely.

It is not a false person, but a fabricated person

Let us begin with the word, because almost everything is decided in it. To call Miquela “false” closes the question before posing it, because it assumes that the person is a fact —something one has or does not have— and that this image, by not having it, lies. To call her “fabricated” opens it, because the person stops being a given and becomes a result, something that is produced, granted and sustained.

The intuition is not new in anthropology. Almost a century ago, Marcel Mauss showed that the “person” —that thing we tend to take as immediate self-evidence, an individuality with an interior, a name, a biography and moral responsibility— is not a universal of the species but a historically constructed form. The word itself comes from persona, the mask of the Latin theatre. First the role, then the face; first the place one occupies in a social order, and only much later the inner self we believe we find beneath it today. Not all societies have distributed that standing in the same way, nor granted it to the same beings. The person is not discovered, it is invested. Whoever receives the mask can speak, be named, answer for their acts and receive reproach. Whoever does not receive it, cannot.

If the condition of being a person is something a culture grants, rather than something nature hands over, the question about Miquela changes shape. It stops being “is she really a person?” —which admits only a yes or a no, and a dull no at that— and becomes “under what conditions is she granted, in practice, the place of a person?”. The indignation of that December was, seen this way, a granting in act. To reproach her for something, those who reacted had already handed her the role. They treated her as the occupant of a position from which one can offend, lie, appropriate. They had put the mask on her without her deciding to wear it.

There remains, then, the question of how. Because a mask does not hold itself up. Someone carves it, someone wears it, someone looks at it and decides to believe. A fabricated person is not an object but an assembly —a web of hands, narratives, brands, platforms and audiences— that is worth taking apart piece by piece.

The image that acts

If a fabricated person is an assembly, the first piece is the image itself. And it is worth resisting the temptation to see it as an ornament —a shell behind which “what really matters”, the human team, would lie. Alfred Gell proposed thinking of certain images not as signs that represent something but as indexes from which the viewer infers a will. Before the idol, before the portrait, before the face that looks at us from the screen, we do not read only “this signifies a person”; we end up attributing intention, desire, a gaze to it. The image does not describe agency, it captures it. Miquela’s face works in just this way. It is not the illustration of a character who exists elsewhere; it is the point from which those who look at her deduce that there is someone feeling something, and from which they grant —or withdraw— the mask.

But the image, on its own, is not enough either. Behind the “I” that speaks in the first person there is no closed individuality but a cast. Marilyn Strathern, in very different ethnographic contexts, showed that the person can be thought of not as an indivisible unit but as a partible being, made of the relations that compose it. A synthetic person is such almost literally. Her voice is written by a scriptwriting team, her face is modelled by a studio, her biography is decided by a company, her reach is governed by an algorithm and her credibility is granted by those who follow her. Miquela’s “I” does not hide a person behind it; it is distributed among all those hands at once. There is no single ventriloquist hand, but a distributed person.

Hence the tempting question —“who is really behind it?”— falls short. Reducing the phenomenon to hidden human agency —a community-management profile with a script— explains it as badly as reducing it to the machine —pixels and a model. Bruno Latour taught us not to choose between those two shortcuts. In a network, persons, objects, texts, platforms and devices can act as actants, and agency does not appear as a property someone possesses in advance but as an effect the network produces. The image does something, the script does something, the platform does something, the audience does something; and “Miquela”, as a person capable of offending or of selling, is the result of all that joint doing, not its first cause.

This gives a concrete way of looking. Before any fabricated person one can ask, piece by piece, who speaks for her, who designs her, who is paid through her, who believes in her, with whom she relates and what body is built for her. These are six questions, not one, and they rarely have the same answer —nor the same weight. In some cases the centre of gravity lies in the voice; in others, in the body; in others, in the money. The three figures that follow do not illustrate the same thing three times. Each one strains a different piece of the assembly and, read together, they let the whole factory be seen.

Miquela, the voice that invents a past

Of the three figures, Miquela is the one who most clearly speaks. She does not merely pose. She has opinions, moods, a circle of friends, fallings-out, a musical career, an origin story. She appeared in 2016 as a young Angeleno of Latina descent, presented again and again on the threshold of nineteen, and ever since her age seems to belong less to the calendar than to the script. That detail, which seems anecdotal, is the first clue to how she is made. A person has an age that advances; Miquela has an age that is decided.

What sustains her is not the image but the narrative. A computer-generated face, however accomplished, is not enough for someone to be treated as a person. A biography is needed, a before, an origin, ties, contradictions, the promise of an after. Miquela has all of that, and all of it is written. Her ancestry, her activism, her friendships with real people and with other synthetic figures, her public break-ups and reconciliations. Each piece is calculated to make an interior legible. Here it is worth recalling Kopytoff and the idea that commodities have a cultural biography —that an object acquires value and meaning through the history attributed to it. The synthetic person takes that logic one step further. Her biography is not reconstructed over time, like that of an object passing from hand to hand; it is designed from the very first day, backwards and forwards, the way the past of a fictional character is written so that the present will seem credible.

And yet what is decisive is that this fabricated biography produces effects that are not fictional. When Miquela “speaks” of her identity, her ancestry or a painful experience, we do not receive it as we read a character in a novel —at a distance, knowing it is paper. We receive it as we receive someone who confides in us, with empathy, with suspicion and with judgement. The first-person voice, sustained over time and accumulated post by post, does something the face on its own would not do. It turns an image into an interlocutor. And whoever occupies that place can be believed, contradicted, defended or reproached. The indignation of 2019 was possible because, by then, years of narrative had installed Miquela in the place of someone who has a life to tell. The moral reproach was not an anomaly. It was the bill for a well-built biography.

It is worth keeping in view who pays and who is paid by that construction. Miquela’s intimate voice —confessional, vulnerable, close— is precisely the register that sells best, and not by chance. Closeness is the asset. The more she is treated as a person with an interior, the more effective she is as a commercial vehicle, because the trust granted to her transfers without friction to the brands around her. Here the voice is not the opposite of business but its most finely tuned form. But that crossing —that of the person who is also a commodity— is already the problem of the third case, and it is worth letting it ripen. First, a figure whose centre of gravity is not the voice but the body, and with the body, race and the question of who has the right to fabricate it.

Shudu, the body that another authorship fabricates

If in Miquela the centre of gravity is the voice, in Shudu there is barely a voice. She does not confide, she does not opine and she hardly narrates. She poses. Shudu is, above all, a body and a face —luminous dark skin, features of impossible symmetry, sometimes the iindzila, the neck rings of the Ndebele people of South Africa. The fabrication, here, is not spread across a biography that accumulates post by post; it is concentrated in the image. And the image, in her case, does one very precise thing. It offers itself as black beauty.

For a time it worked exactly like that. Before it was known what she was, Shudu’s photos circulated through accounts that celebrate black women, with tags such as #blackgirlsrock or #melanin; she was admired like a dazzling dark-skinned model. It is Gell’s mechanism in its pure state. The image did not report that “this represents a black woman”, it directly provoked admiration, identification, pride. It captured a very particular racial and affective response. Later it became known that behind her there was no black model but an external white authorship. Cameron-James Wilson, a British person devoted to fashion photography, who had modelled her in 3D drawing inspiration, among other things, from a Barbie called “Princess of South Africa”.

There lies the controversy, and it is worth framing it in its serious version, not the one of a simple “she is fake”. What Lauren Michele Jackson wrote in The New Yorker —that Shudu was the digital projection, made by a white man, of a real black womanhood— is not a complaint about deception but about structure. A black woman’s body fabricated, exhibited and monetised by someone who controls it from outside. The criticism inscribed it within a known genealogy. Read through Patricia Hill Collins, the figure appears as the image of a black woman devised and controlled from outside, and Lauren Michele Jackson, in The New Yorker, read it in continuity with a longer history of racial appropriation. The blackface of the minstrel shows, which —according to Eric Lott— allowed white audiences to enjoy their fascination with blackness without having to deal with black people. The underlying reproach is not aesthetic. Real black women fight for the opportunities granted here, without friction, to an image that none of them inhabits. “We don’t want to be a trend”, Naomi Campbell had said of black models; Shudu turned the phrase uncomfortably literal.

Wilson has defended himself by saying that this is art, not exploitation, and that his creature does not replace human models but accompanies them. Perhaps that was his intention. But the anthropological question is not about the intention of a particular person but about the form of the assembly. And Shudu’s assembly separates, in a way that Miquela’s did not, two things we usually assume to be joined in a person. The body that is seen and the hand that produces it and is paid for it. The fabricated person and the fabricated identity do not coincide. They belong to different people. What is more, not even the body is wholly synthetic. The project itself has also taken in digitised real bodies —models turned into avatars or set to circulate alongside Shudu—, so that the “distributed person” we spoke of here includes something more than code and image. It includes living black bodies whose presence can be absorbed by the fiction that surrounds them.

Miquela concentrated the fabrication in the voice; Shudu, in the body. Both, even so, still ask to be treated as someone. The third figure renounces even that. She never pretended to be a private person, she was a brand from the very first day. And even so she ended up occupying the place of one.

Lu do Magalu, the brand that became a person

Miquela hid the factory behind a private person; Shudu, behind a body. Lu hides nothing. She was born in 2003 with not the slightest pretension to intimacy. She was, literally, the voice of a shop. Magazine Luiza needed a face for its e-commerce in a country where buying online still bred mistrust, and Lu was that, a virtual assistant who explained products and accompanied those buying for the first time on the web. A logo with a face. Not a person pretending to be one, but a commercial instrument that never concealed being one.

And yet, two decades later, she occupies the place of a public person. She has tastes, moods, a “daily life” —trips, workouts, real and virtual friendships—, opinions, memes, musical enthusiasms. She is followed, quoted, loved. What is decisive is how that transition happened, because it was no accident. It was a strategy, and the company names it without embarrassment. Her influence, say her own management teams, began by “humanising” her. That is, the person did not appear in spite of the commercial function but as its most finely tuned form. First she was useful —a whole decade generating trust, not attention—; then she became close; and closeness, turned into personality, proved to be the best sales vehicle the brand had ever had.

This is where “person” and “commodity” cease to be opposites. We recalled with Kopytoff that commodities have a biography. Lu’s case shows the crossing in its extreme form. A commodity for which an interior is fabricated because the interior sells. The trust granted to someone with a life of her own transfers, without friction, to every product that figure recommends. And like any public person, Lu does something a catalogue could not. She takes sides. A retailer’s mascot that comes out in favour of feminism and gets involved in social causes is the clearest case of a brand occupying the place of a moral subject, with an audience that grants it to her. The indignation that in Miquela was reproach is, in Lu, adherence. Whoever admires Magalu follows Lu and, in following her, grants her the standing of someone one can agree with.

The piece of scale is missing, and it is worth handling with care, because it is the most opaque. Lu is billed as the biggest virtual influencer in the world, with tens of millions of people following her across platforms. But the concrete figures dance according to the source —and according to who gives them—, they are openly presented as estimates, and the most inflated version in circulation does not match the one given by the soberer counts. I will not fix a number, and the reason is not only editorial prudence. The opacity itself is part of what Lu is. A commercial instrument whose magnitude depends on scattered, shifting counts repackaged by platforms, brands and industry sources; her power is measured in an aggregate metric that never quite settles for anyone observing it from outside. The fabricated person culminates here in a person whose reach, like her biography, forms part of the infrastructure that sustains her.

Three figures, three pieces of the same assembly. The voice that invents a past, the body that another produces, the brand that becomes a person. Separately, each one strains a different nerve —biography, race, value. Together they pose a single question, which is the one that remains. If a person can be fabricated like this, what exactly is it that this factory produces?

What the factory produces

A fabricated person does not produce a false person. It produces effects, and they are real. It is worth seeing them for what they are, not loose chapters but consequences of one and the same gesture. Four are worth dwelling on.

The first thing it produces is authenticity, and here an intuition must be undone. We tend to think that these figures lack authenticity and fake it. But authenticity is not a property one has or does not have. It is something that is produced, and these figures produce it effectively. The interesting question is not only whether they deceive. Sometimes the fabrication is hidden; at other times it is in plain sight. What is decisive is how an audience relates to something it knows, at some level, to be fabricated. The answer has an old shape, which Octave Mannoni summed up in a formula, “I know very well, but all the same”. We are not talking about gullible people who have been hoodwinked; they know perfectly well what they are looking at, and even so they treat it as someone. Knowledge and belief do not cancel each other out. They coexist. On that ground works what Kim calls meta-authenticity. The authenticity of a synthetic influencer is not a prior given but the dynamic result of coordinated performances between human and non-human agents. And an uncomfortable finding shows how far the effect reaches. In an online experiment, some of these figures were perceived as more authentic than real people, above all among those who most strongly activated the machine heuristic, that disposition to attribute consistency or reliability to what is produced by technical systems. Not less authentic. More.

The second thing it produces is value, and of a peculiar kind. The figure is at once commodity and person, without contradiction. Marx described the fetish of the commodity as that object which seems to have a life of its own; the synthetic person takes the image to the letter, because its life of its own is exactly what is sold. The affection granted to it is not the opposite of business but its most profitable form, because the trust placed in someone with an interior transfers without friction to whatever that figure touches. The person-effect does not accompany value, it generates it.

The third thing it produces is a body, and with it a desire. Not just any body, but one with no age, no tiredness, none of the “imperfection” that gives away the living, always available and —in the vast majority of cases— female and often sexualised. Le Breton showed that the body is not a brute biological fact but a cultural construction; these figures build one freed from the human condition. Communication studies record it. Female virtual influencers tend to reinforce unattainable ideals of youth, beauty and explicit sexuality, at no cost to themselves, because they neither age nor err. But the datum must be read in an anthropological key. A body that cannot decay is not a better body; it is a body exempted from being mortal. And that exemption is exactly what is offered and what is desired.

There is, moreover, a feminist question that ought not to be left in a footnote. This essay works with three figures coded as feminine, and that coincidence should not be read as chance. Without a census of our own we cannot assert a global proportion, because there are also male avatars, virtual bands, VTubers and non-feminised corporate characters. But in the commercial circuit of fashion, beauty, lifestyle and affective attention, the synthetic person frequently appears as a feminised body, available to be looked at, dressed, corrected, sexualised or made trustworthy. The question is not only whether these figures represent women, but what cultural work their feminisation does, what desires it organises, what bodies it displaces, what bodies it erases and who captures the value of that availability.

The fourth thing it produces is bond and, symmetrically, rejection. People grow fond, defend, identify, take offence. The bond is real on the human side, even if on the other side there is no singular interior able to reciprocate it. It is a sociality with a distributed counterpart, affection and judgement directed at a being that responds only through the network that sustains it. The indignation with Miquela, the controversy of Shudu and the adherence to Lu are three forms of one and the same fact. These figures produce genuine social relation without a genuine subject on the other side.

Authenticity, value, body and bond. None is a side effect; all four are the product. What the factory delivers is not a fake person but the social yields of a person —credibility, money, desire, attachment— detached from any interior to back them. And there lies what is worth thinking through to the end.

What the factory teaches us

It is worth saying, at last, what we have been looking at. Not false persons. That verdict closes the matter at the worst moment, just when it begins to be interesting. What we have seen are techniques —cultural, not only technical in the computing sense— for fabricating the three things we usually believe only a real person can give —presence, trust and value. Miquela fabricates presence with a voice and a past; Shudu, with a body; Lu, with two decades of usefulness turned into closeness. And all three pull it off.

Hence the thesis this essay defends, the one that makes sense of treating artificial intelligence not as a tool but as an anthropological object. These figures do not, at bottom, teach us much about machines. They teach us about ourselves, about the operations by which a collectivity grants something the standing of a person. They work as a reagent, in the chemical sense —a substance added not to matter in itself but to make something else visible. What they reveal is that the condition of being a person, that thing we took for immediate self-evidence, lets itself be dismantled into pieces —biography, body, voice, moral position, value, affection— and that those pieces can be assembled, one by one, with no interior beneath to sustain them.

We return thus, by the long road, to where we began. Marcel Mauss held that the person is not discovered, it is invested. The synthetic person is the limit case that makes that thesis hard to dodge. Here the investiture occurs without a lived biography preceding it, and even so it works. Those who reacted to Miquela were not mistaken to reproach her; they granted her, in act, the place from which reproach can be received. The mask —persona, as we recalled— held up with no face behind it.

And here is what is worth thinking through to the end, because it is the uncomfortable part. The unease these figures produce does not come from their being radically different from us. It comes from the opposite. From their not being so different after all. The operations we see in them in their pure state —a biography that is written, a body that is cared for and exhibited, a voice that is fine-tuned, a trust that is turned into value— also operate, more slowly and more disguised, in the fabrication of any public person. Perhaps of any person. The difference is not that they are fabricated and human public persons are not. It is that in them the seams show.

There is no need, to finish, for any verdict about the future. Neither the alarm that machines will replace us, nor the enthusiasm that they will free us. The question was anthropological and remains so. What these fabricated persons put on the table is not their own falsity but an old truth we would rather not look at head-on. The person —the presence we recognise, the trust we grant, the value we let circulate— has always been, to some degree, a collective production. These figures do not lie to us about what they are. They tell us, without meaning to, something about what we are.

References