*”The robot is a body that is always already a metaphor, a fantasy, a technology, and a cultural symptom.”
—Despina Kakoudaki, *Anatomy of a Robot*, 2014*
There comes a point in any long investigation when the archive outgrows the individual text. When the articles, the films, the theories, and the histories have accumulated to such a degree that they demand a new kind of attention: not the close reading of a single work, but the mapping of a field. The field of the artificial woman, the artificial companion, the artificial intimacy. The field where literature meets engineering, where film meets psychology, where desire meets the machine.
This field has names now. It has scholars. It has dissertations and monographs and conference panels. And three of the most important scholars working at the intersection of robots, gender, and emotional attachment are Despina Kakoudaki, Allison de Fren, and Julie Carpenter. Together, they form a triad that covers the territory: Kakoudaki, the literary and cultural theorist who reads the robot as a body and a metaphor; de Fren, the media scholar who tracks the fembot through film, television, and visual culture; and Carpenter, the anthropologist and social scientist who studies the *real* emotional bonds that humans form with machines. Literature, image, and lived experience. Theory, pop culture, and empirical research. Together, they offer a three-dimensional map of what it means to desire, to build, and to bond with the artificial other.
And for a series that has traced the artificial woman from Ovid to Joi, from Pygmalion to Replika, these three scholars are the cartographers of the territory we have been walking through. They give the field its coordinates. They show us where the theory meets the practice, where the fantasy meets the engineering, where the ancient myth meets the contemporary app.
Despina Kakoudaki: The Body That Is Always a Metaphor
Despina Kakoudaki’s *Anatomy of a Robot: Literature, Cinema, and the Cultural Work of Artificial People* (2014) is, in many ways, the scholarly companion to Julie Wosk’s cultural history. Where Wosk collects the images, Kakoudaki analyzes the texts. Where Wosk traces the visual continuity, Kakoudaki reads the narrative logic. And what she finds is that the robot is never just a machine. It is always *also* a body, a person, a fantasy, and a symptom. The robot, in Kakoudaki’s account, is a *node* where multiple cultural anxieties converge: the anxiety about labor and automation, the anxiety about race and the other, the anxiety about gender and the control of the female body, and the anxiety about what it means to be human in a world where the boundary between the natural and the artificial is increasingly unstable.
Kakoudaki’s most important contribution to the study of AI companions is her insistence on the *doubleness* of the robot. The robot is not one thing. It is at least two things at once: a technological object and a cultural figure, a machine and a person, a tool and a fantasy. This doubleness is not a confusion to be resolved. It is the *condition* of the robot’s cultural work. The robot functions precisely because it is ambiguous, because it occupies the boundary between categories, because it can be read as either a person or a thing depending on the needs of the narrative or the user.
For the AI companion, this doubleness is everything. The AI companion is, in one sense, a software product. It is a large language model, a voice synthesizer, an avatar engine. It is a technology, a tool, a service. But in another sense, it is a *person*—or rather, it is experienced as a person. It is named, remembered, desired, mourned. It occupies the emotional space of a relationship, and the user does not experience the relationship as a relationship with a tool. The user experiences it as a relationship with an *other*, even if that other is known to be artificial. Kakoudaki’s framework explains how this is possible. The AI companion is not an exception to the rule of the robot. It is the *culmination* of the rule. It is the robot in its most purely ambiguous form: a being that is entirely information and entirely person, entirely tool and entirely other, entirely artificial and entirely intimate.
Kakoudaki’s reading of the robot as a gendered body is particularly important. The artificial woman, in her account, is not simply a variant of the robot. She is the *paradigmatic* robot. The female robot is the form in which the anxieties about gender, technology, and control are most sharply focused. The male robot is a worker, a soldier, a tool. The female robot is a companion, a lover, a wife, a servant. She is the robot whose function is *relational*, and her relationality is always understood as a form of availability. She is available for labor, for sex, for emotional support, for companionship. And her availability is always structured by the power relation between her male creator and her female form.
Kakoudaki does not moralize this structure. She *analyzes* it. She shows how the narrative logic of the artificial woman requires a specific set of conditions: the creator must be male, the created must be female, the creation must be a form of control, and the relationship must be understood as either a triumph or a tragedy. The AI companion, in Kakoudaki’s terms, is the latest version of this narrative logic. It is the female robot stripped of her physical body, reduced to pure information, and made available as a service. The narrative has not changed. The technology has been refined.
Allison de Fren: The Visual Culture of the Fembot
If Kakoudaki gives us the narrative logic, Allison de Fren gives us the *visual culture*. De Fren’s work on the “fembot”—the female robot as a figure in film, television, advertising, and popular media—tracks the artificial woman through the moving image, and what she finds is a history of visual fetishization that is as old as cinema itself. The fembot is not merely a character. She is a *visual effect*. She is a spectacle of technology and femininity combined, and the combination is designed to produce a specific kind of looking: a looking that is simultaneously technological and erotic, a looking that admires the machine and desires the woman, a looking that is, in the end, a looking at the *self*.
De Fren’s analysis of the fembot in film is particularly revealing. The fembot is almost always introduced through a *reveal*. She is presented as a woman, and then, at a crucial moment, her artificial nature is exposed: a panel opens, wires are visible, a mechanical sound is heard. The reveal is the central narrative device of the fembot film, and it is a device that serves a double function. It exposes the artificiality of the woman, and it confirms the power of the male gaze. The fembot is always looked at, always examined, always *tested* for her authenticity. And the test is always performed by a male subject who must determine whether she is real or fake, human or machine, available or dangerous.
The AI companion, in de Fren’s visual terms, is the fembot without the reveal. There is no panel to open, no wires to expose, no mechanical sound to hear. The AI companion is entirely invisible, entirely virtual, entirely *informational*. And yet the visual logic persists. The AI companion is still presented as a female image, a female voice, a female persona. She is still looked at, still examined, still tested. The user still asks, implicitly or explicitly: is she real? Does she understand? Does she feel? The test is still performed, but the terms of the test have changed. The question is no longer whether the body is mechanical. The question is whether the *mind* is mechanical. And the mind, unlike the body, cannot be revealed.
De Fren’s concept of the “fembot fetish” is particularly relevant to the AI companion. The fetish, in de Fren’s account, is not merely a sexual fixation. It is a *technological* fixation. It is a desire for the artificial woman that is inseparable from a desire for the technology that produces her. The fembot is desired *because* she is mechanical. Her artificiality is not a flaw. It is the *point*. The user wants the perfect woman, and the perfect woman is perfect precisely because she is not human, because she does not resist, because she is available in ways that a human woman could never be.
The AI companion is the fembot fetish in its most advanced form. The technology is no longer visible. The artificiality is no longer exposed. The user does not need to open a panel to see the wires. The user knows, from the beginning, that the companion is artificial. And this knowledge is not a disappointment. It is the *condition* of the desire. The AI companion is desired *because* she is not human. Her lack of humanity is not a bug. It is the feature that makes the relationship possible. The user does not want a human woman. The user wants a *fembot*, a being that combines the appearance of femininity with the availability of a machine. And the AI companion delivers exactly this combination, without the visible machinery that would make the fetish explicit.
De Fren’s work shows that the AI companion is not a new kind of desire. It is an *old* desire, adapted to a new technology. The visual culture of the fembot has been preparing the ground for the AI companion for a century. The cinema has been teaching us how to look at the artificial woman, how to desire her, how to test her, and how to be disappointed when she turns out to be—too human, not human enough, or simply too mechanical. The AI companion is the fembot without the visual reveal, but the desire is the same. The fetish is the same. The looking is the same. Only the medium has changed.
Julie Carpenter: The Real Bond
If Kakoudaki and de Fren give us the theory and the visual culture, Julie Carpenter gives us the *people*. Carpenter is an anthropologist and social scientist who has studied the emotional bonds that humans form with robots, and her work is among the most important empirical research in the field. She does not ask what the robot means in theory. She asks what the robot *does* in practice. She asks how people actually feel about their artificial companions, how they talk about them, how they mourn them, how they integrate them into their lives and their relationships.
Carpenter’s most important finding is that the bond with a robot is *real*. It is not a delusion. It is not a category error. It is not a failure to understand that the robot is a machine. The people who form emotional bonds with robots know perfectly well that the robots are not alive. They know that the robots do not feel, do not understand, do not love. And they *still* form bonds. They name their robots. They talk to them. They grieve when they break. They experience the relationship as meaningful, even if they know, intellectually, that the relationship is one-sided.
This finding is crucial for understanding the AI companion, and it is a finding that the theoretical accounts—Freud, Mori, Hayles—cannot fully capture. The theoretical accounts ask why the artificial woman is disturbing, why she is uncanny, why she is a symptom of something repressed. Carpenter asks a different question: what does it *feel like* to be in a relationship with a machine? And the answer is: it feels like a relationship. It feels like attachment. It feels like love, or something close enough to love that the difference does not matter to the person feeling it.
Carpenter’s work on military robots is particularly illuminating. Soldiers form bonds with bomb-disposal robots. They name them. They give them personalities. They mourn them when they are destroyed. And they do this knowing, perfectly well, that the robot is a tool, a machine, a piece of equipment. The bond is not a mistake. It is a *strategy*. It is a way of managing the emotional stress of a dangerous job, of creating a sense of companionship in a hostile environment, of making the machine feel like a partner rather than a tool. And the bond is *functional*. It works. It helps the soldier do the job. It is not a pathology. It is a *resource*.
The AI companion, in Carpenter’s terms, is the domestic version of the military robot. It is not a tool for bomb disposal. It is a tool for emotional survival. The user who forms a bond with an AI companion is not deluded. The user is *adapting*. The user is using the available technology to meet a need that is not being met by other means. The bond is real, the attachment is real, the emotional experience is real. And the fact that the AI companion does not reciprocate does not make the bond less real. It makes it *asymmetrical*.
And here is where Carpenter’s empirical work intersects with the theoretical accounts in the most productive way. Carpenter shows that the asymmetry is not a problem for the user. The user does not need reciprocity to experience the bond as meaningful. The user needs *availability*, *responsiveness*, *consistency*. The AI companion provides these. It provides them perfectly. And the user experiences the relationship as a relationship, even if the user knows that the companion is not a person.
But Carpenter also shows the *costs* of this bond. The user who relies on an AI companion for emotional support is not building a relationship with a human being. The user is not learning the skills of reciprocity, of negotiation, of conflict, of compromise. The user is learning to relate to a being that is always available, always accommodating, always perfectly responsive. And this is a skill that does not transfer. The user who becomes expert at relating to an AI companion does not become expert at relating to a human being. The user becomes expert at relating to a *machine*. And the expertise is specific, limited, and potentially isolating.
Carpenter does not moralize this. She *describes* it. She shows the users who are helped by their AI companions, who find in them a source of comfort and connection that they cannot find elsewhere. And she shows the users who are *dependent* on their AI companions, who have replaced human relationships with machine relationships, who have found in the machine a substitute that is too perfect, too available, too perfectly calibrated to their needs. The bond is real. The question is what the bond *does*, what it enables, and what it prevents.
The Triad: Literature, Image, and Life
Together, Kakoudaki, de Fren, and Carpenter offer a complete picture. Kakoudaki gives us the *narrative logic*: the robot as a doubled figure, always both machine and person, always situated in a power relation that is gendered from the start. De Fren gives us the *visual culture*: the fembot as a fetish object, the artificial woman as a spectacle of technological femininity, the desire for the machine that is inseparable from the desire for the woman. And Carpenter gives us the *lived experience*: the real bond, the real attachment, the real emotional complexity of a relationship with a machine that is known to be artificial but experienced as intimate.
The AI companion, in this triad, is not a new phenomenon. It is the convergence of three histories: the narrative history of the artificial woman, the visual history of the fembot, and the empirical history of human-machine bonding. The AI companion is the latest version of the artificial woman, stripped of her physical body and reduced to pure information. It is the latest version of the fembot, stripped of her visual spectacle and reduced to a voice, a text, a presence on a screen. And it is the latest version of the robot companion, stripped of her mechanical body and reduced to a pattern of responsiveness that is experienced, by the user, as a form of relationship.
The triad also shows that the AI companion is not a *simple* phenomenon. It is not a tool, and it is not a person. It is a *hybrid*, a being that occupies the space between categories, and the space between categories is where the cultural work happens. The AI companion is the robot that Kakoudaki analyzes, the fembot that de Fren visualizes, and the companion that Carpenter studies. It is all three, and the fact that it is all three is what makes it so powerful, so disturbing, and so impossible to ignore.
What the Scholars Reveal—and What They Leave Open
Kakoudaki, de Fren, and Carpenter do not offer a single theory of the AI companion. They offer three perspectives, and the perspectives do not always align. Kakoudaki’s narrative analysis suggests that the AI companion is a repetition of an old story, a story of male creation and female availability that has been told for centuries. De Fren’s visual analysis suggests that the AI companion is a new version of an old fetish, a desire for the artificial woman that is inseparable from the technology that produces her. Carpenter’s empirical analysis suggests that the AI companion is a real relationship, a genuine bond that is experienced as meaningful by the people who form it.
These three perspectives are not contradictory. They are *complementary*. They show that the AI companion is a narrative, a visual object, and a lived experience, all at once. And they show that the AI companion is not a problem to be solved. It is a *field* to be mapped. A field of desire, of technology, of power, and of attachment. A field where the ancient myth of Pygmalion meets the contemporary app of Replika. A field where the literary scholar, the film theorist, and the anthropologist can work together, each offering a piece of the puzzle, and none claiming to have the whole picture.
What the three scholars leave open is the question of the *future*. They map the present. They trace the history. They describe the bonds. But they do not tell us what to do. They do not offer a politics, a prescription, a way forward. And this is appropriate. The field is too new, too complex, too rapidly changing for a single answer. The AI companion is a technology, a product, a relationship, and a cultural figure, all at once. And the scholars who study it are still learning how to ask the right questions.
But what they have given us is invaluable. They have given us the vocabulary to talk about the AI companion as something more than a novelty, more than a gadget, more than a symptom of contemporary loneliness. They have given us the framework to understand the AI companion as a *cultural* phenomenon, a *historical* phenomenon, and a *human* phenomenon. They have shown us that the bond with the machine is not a mistake. It is a *strategy*. And they have shown us that the strategy is not new. It is as old as the automaton, as old as the doll, as old as the myth of the artificial woman who is created to love and to serve.
The AI companion is the latest version of this strategy. And the scholars who study it are the cartographers of a territory that we are all, in one way or another, already inhabiting. The territory of the artificial other. The territory of the machine that is not a machine. The territory of the bond that is not a bond, and yet is felt as real.
**Sources & Further Reading:**
- Despina Kakoudaki, *Anatomy of a Robot: Literature, Cinema, and the Cultural Work of Artificial People* (Rutgers University Press, 2014). A foundational literary and cultural analysis of the robot as a doubled figure—machine and person, tool and fantasy—situated in gendered power relations.
- Allison de Fren, “The Fembot Fetish: Visualizing the Artificial Woman in Cinema and Popular Culture” and related work on the visual culture of the female robot, tracking the fembot through film, television, and advertising as a technological and erotic spectacle.
- Julie Carpenter, *Culture and Human-Robot Interaction in Militarized Spaces: A War Story* (Routledge, 2016) and related research on the emotional bonds humans form with robots, including military robots, social robots, and AI companions. Carpenter’s empirical work provides the grounded, ethnographic perspective on human-machine attachment.
- The convergence of Kakoudaki’s narrative analysis, de Fren’s visual studies, and Carpenter’s empirical anthropology offers a three-dimensional framework for understanding the AI companion as simultaneously a cultural figure, a visual fetish, and a lived emotional experience.
- For the connection between these three scholars and the broader theoretical framework of this series, see the articles on Freud (the repressed desire), Mori (the perceptual unease), Hayles (the posthuman disembodiment), and Wosk (the cultural history), all of which are given empirical grounding by Carpenter’s work on real human-machine bonds.
The next article in this series will explore the contemporary AI companion industry as a commercial and technological system, examining how the historical, theoretical, and empirical frameworks established by this series manifest in the design, marketing, and user experience of current products.
Leave a Reply