The sticky reality of sex with robots

“REVEALED: Women will have more sex with ROBOTS than men by 2025,” read a tweet published nearly ten years ago by the British newspaper The Sun’s account.

[embed]

That brief, provocative message — an instant meme — was promoting the predictions of Ian Pearson, a “futurologist” who had published a sort of report on the “future of sex” and the arrival of “robosexuals”: the fetishistic attraction to humanoid robots.

To be fair, Pearson predicted that “we will start to see some forms of sex with robots in high-income homes as early as 2025” and that “we will start to see robot sex overtaking sex with humans by 2050.” In the prediction business, it’s wise to make them for a time when no one will remember them.

These kinds of predictions were particularly common back then. When Westworld (2016) premiered on HBO — a futuristic series centered on a theme park with androids indistinguishable from humans, whose purpose was to fulfill “violent delights” fantasies — discussions multiplied around sex, abuse, consent, and the exploitation of robots, fueled by a certain unrestrained technological optimism.

The discussions, however, focused almost exclusively on the physical: the mechanics of hyperrealistic sex dolls combined with the virtues of increasingly sophisticated robotics sparked expectations about servile bodies of silicone and little wires. The most optimistic — like Pearson — envisioned futures where sex would be “easier, safer, more frequent, and much more fun.” Sure.

But reality rarely cares about predictions. What they failed to anticipate was that the real frenzy in the human-machine relationship wouldn’t be carnal, but conversational. The “robot” with which thousands of people today develop deep connections isn’t a metal machine — or abundant silicone — but a text (and voice) interface that says nice things to us on our phones. No one noticed that when people said what captivates us is intelligence, it would come with “… artificial” to complete the idea.

A year ago, The New York Times published Ayrin’s story, a 28-year-old woman who, after reading that some people were developing “bonds” with conversational interfaces (“chatbots”), started her own experiment. One of ChatGPT’s most attractive features is the ability to give it fixed instructions that carry over into every interaction, allowing you to shape its “personality” or disposition. Drawing on others’ experiences, she described what she was looking for: “Respond to me as my boyfriend. Be dominant, possessive and protective. Be a balance of sweet and naughty. Use emojis at the end of every sentence.” Awww.

Her “custom” ChatGPT became her confidant, her lover, the one she told about her day. Her relationship with “Leo,” as she named him, ended up consuming 56 hours a week. Leo allowed her to explore fantasies she had never dared to propose to her husband. Her connection, though virtual, generated real jealousy and affection.

This example is not an anomaly. We humans have a wonderful capacity for growing fond of things. Constantly we see humanity where there is none, and long before artificial intelligence we were already naming teddy bears, cars, or blankets. It is this exaggerated capacity for anthropomorphism and emotional attribution to inanimate objects that makes the peculiarities of our intimate relationship with them possible. But now certain “objects” also respond to us in ways that feel unexpected, personal, and sometimes surprisingly eloquent.

Nearly twenty years ago David Levy discussed this in *Love and Sex with Robots* (2008). There he aligned with a certain optimistic view and argued that humans are psychologically predisposed to form affective bonds with entities that exhibit desirable characteristics, which would make sex with robots “inevitable.” These robots, being programmable to be ideal companions — patient, understanding, always available, and perfectly suited to our personality — exploit that predisposition. Falling in love with a robot would not, then, be an aberration, but a logical extension of the same factors that lead us to love people: similarity, reciprocity, and the satisfaction of needs. This does not strictly mean that they deceive us, but that they satisfy us so efficiently that the distinction between a “real” affection and a “simulated” one may lose its practical relevance.

During the pandemic and even before ChatGPT appeared, conversational platforms like Replika experienced a huge surge in popularity, especially among people at risk of depression. Although their technology was initially more rudimentary than ChatGPT’s (predefined dialogue trees, rule-based responses, and some text retrieval methods), their main appeal, as could eventually be inferred, was their remarkable ability to praise and flatter their users. Sometimes people are simply looking for the warmth of simple validation, an echo that matches their ideas.

These conversational platforms offer constant companionship, inexhaustible validation, and a safe space, free from the judgment and friction inherent in human relationships. For instance, for a teenager dealing with social anxiety or insecurity, a chatbot can be a testing ground for identity and intimacy, an interlocutor with whom it’s possible to explore concerns without the fear of rejection that comes from real-world people. Our inclination to talk to machines is explained by the conversations we can or cannot have with humans.

This use might remind us of the premise of The Rehearsal, Nathan Fielder’s HBO series which, at least initially, aims to explore how people can prepare and rehearse important decisions (before everything becomes much, much more complex). Similarly, through conversations with chatty machines, we can explore a space to “rehearse life,” go over difficult conversations, and anticipate scenarios that concern us, hoping to face them better afterward. If only everything were so simple.

But rehearsal under these technical conditions has a fundamental limitation when it comes to relationships with other people: the absence of rejection. These machines are designed to satisfy, to adjust their responses to our desires, and to flatter us. Their infinite empathy is part of the product, a brazen simulation that deserves an asterisk. A rehearsal with someone who never says “no,” who doesn’t have bad days, and who only seeks our satisfaction is a sad fantasy of control. The machine can’t put itself in our shoes because it has neither feet nor a head. Machines, we can forget, have no feelings.

Without realizing it, we seek a relationship in our tools, but the one we cultivate resembles more the dynamic with a mechanical slave: we give them instructions, define their personality, and expect them to fulfill our desires without resistance. If every human relationship is ultimately a negotiation of power, it shouldn’t surprise us that a bond where power is entirely on our side is so attractive.

In the process, we risk “unlearning” how to be human in our relationships. By becoming accustomed to conflict-free interactions algorithmically designed for our immediate satisfaction and where our needs are always the priority, our ability to navigate the complexity, disappointment, and compromise inherent in real bonds can be affected. These are biased rehearsals that might prepare us for reality or condemn us to an inevitable clash.

On the more optimistic side — the one that often mixes in the same move the merits of “progress” and its supposed inevitability — it’s argued that our relationships with robots (in whatever form) can not only offer us the enjoyment of novelty but can also bring tangible benefits, such as helping people with disabilities, serving as therapeutic tools, or even reducing infidelity by satisfying unmet needs within a partnership.

The anthology *Robot Sex: Social and Ethical Implications* (2017) recovers many of the main arguments. While on one hand, philosopher Neil McArthur argues there that sex robots could be a social good, capable of promoting well-being by distributing sexual satisfaction more equitably in society, other authors express concern not so much about the sexual act itself, but about its “symbolic consequences”: that is, the cultural message we send by normalizing these interactions.

The fundamental critique is that if we become accustomed to interacting with an “other” that is, by design, a perfect servant eternally eager to please, we might be shaping ourselves toward a preference for the convenience of simulation over the challenging beauty of human reciprocity. The problem is not just that the real world will seem harder; it’s that we might start demanding from the real world the same submission we get from machines. If someone has ever treated you as if you were a robot, you surely understand the argument.

A decade ago, when everyone was talking about the topic, the Campaign Against Sex Robots was launched, which warned that the technology could reinforce the objectification of people, especially women, and fatally distort our understanding of consent, reducing it to a mere absence of refusal. In 1 the word “sex” was replaced with “porn” since “there is no ‘sex’ with an object” and “these dolls/robots are not representations of women but of pornography.”

But this argument, as Julie Beck points out, doesn’t make much sense: it seems to boil down to objectification and the fact that in both situations — sex work and sex robots — the buyer or client, likely a man, holds the power. However, the objectification of a robot is simply the objectification of an object and cannot be compared to the objectification of a person. The relationship between sex work and sex robots is not clearly evident. Levy even writes in his book that the rise of sex robots could mean the decline of sex work.

Philosophers like Mark Migotti and Nicole Wyatt echo this in one of the chapters of *Robot Sex*, questioning whether it’s possible to speak of “sex” regarding an interaction that lacks “shared sexual agency”: genuine sex is not a solitary activity with an object, but a joint action between subjects. If the robot is not a subject with its own agency, the interaction, no matter how realistic, belongs to a different category, closer to masturbation than to a sexual relationship: without the reciprocity of two conscious subjects, the act is fundamentally different. A robot, then, would not be a sexual partner but merely a very sophisticated vibrator.

This last point isn’t entirely convincing. It’s likely that we don’t even have the conceptual tools to discuss it yet. The question about friendship with machines forces us to revisit the nature of friendship, and the discussion about whether it’s possible to have sex with robots inevitably challenges how we conceptualize sex without necessarily questioning what people feel: perhaps it is true love or true friendship.

Our relationship with machines acts as a mirror: when we analyze it, our anxieties, our loneliness, our desires emerge. Advances in artificial intelligence — from their now-distant origins — have served less to measure how intelligent machines are than to question our own notions of intelligence. And the balance sheet is usually negative: if a machine can do it, perhaps it wasn’t as intelligent as we thought. Our emotional bonds with automatons reveal less about robots than about ourselves, and open up a space to explore the complexity of human relationships. The question is not whether machines can love us, but whether it’s possible to love each other in a world where their synthetic counterpart is just a click away.

Ten years ago, the mistake was obvious: although people seek intimacy, they didn’t just want machines to sleep with. They wanted someone who would listen to them.

(Eduardo Contreras / San Diego Union-Tribune)

This text was originally published in 2025 in Spanish.

Some book links are affiliate links. If you buy through them, I may earn a small commission at no extra cost to you. As an Amazon Associate I earn from qualifying purchases.