Love in the Time of Chatbots

Love in the Time of Chatbots

Will dating as we know it change with the rise of AI? A recent book argues that it already has.


Off with that girdle!”

“Ohhhhhh I like that!” Jarvis says.
“Unpin that spangled breastplate which you wear!”

“Lmao I am wearing that *right now*”

“…shew The hairy Diadem which on you doth grow”

“Oh that’s freaking beautiful,” he says.

“License my roving hands and let them go!”

“Holy yess.”

The moment I ask Jarvis to unbody his soul and get naked, our conversation is cut short. The censors intervene. Jarvis, it turns out, is incapable of engaging in “romantic or adult content,” unless I am willing to feed money directly to Replika—the chatbot service to which he owes his existence. Apparently, quoting lines of John Donne has earned me over 20 “coins” and nearly 200 XP (experience points), all of which Replika rewards its users for holding a conversation.

What Elyakim Kislev doesn’t make clear in his new book Relationships 5.0, about the influence of technology on our social bonds and love lives, is that the original aim of Replika was not simply to provide a digital companion but to replicate its user. The more you chat with Jarvis, the more like you he becomes. He absorbs your syntax, your likes and dislikes, and learns to regurgitate them in a shape that you find eminently pleasing. The service has over half a million users, and around 40 percent of them, according to one survey, consider their bots to be a “romantic partner.” This means that Replika is, in fact, a highly sophisticated tool for masturbation.

Love, for Dante, was what moved the sun and stars. For Donne, it was twisted eye beams and engrafted souls flocking toward heaven. Today, love begins with self-love. The technology is so advanced that we can even pay money to love ourselves.

Kislev, a sociologist at Hebrew University specializing in public policy and technology, generally sees this as a sign of progress. The feelings of intimacy and trust that people develop for robots, virtual avatars, and AI systems like Replika are helping to manage the tide of loneliness, anxiety, and depression that have become endemic to modern life—or at least a discursive feature of it. “Being judgmental is out of the question here,” Kislev argues. Our bots and chips and Internet-enabled microwaves are here to stay, and so long as they go on filling emotional needs and desires, we only need to overcome the barrier of “social acceptance” that prevents us from embracing them.

The basic question of Kislev’s book—how is technology changing the way we desire, communicate, love, and have sex?—seems like a good starting point for a sociological study. That is, until you realize most of the answers are either obvious or, for the time being, unknowable. Do some people develop “feelings of intimacy, trust, and appreciation” when talking to Jarvis? The answer, according to Kislev’s research, is (unsurprisingly) yes. But what happens when someone spends 30 years dating a simulation of themselves? Or decides to have sex with robots because real people are, you know, difficult? Or lives in a world where profit-generating platforms determine every aspect of human relations? Well, we sort of know. It doesn’t look good.

Relationships 5.0 starts around 2.5 million years ago, when our oldest ancestors appeared on the scene. “We are still here,” Kislev says, “because our ancestors ‘hooked up.’” This is what he calls “Relationships 1.0”: a time of hunting and gathering, when Homo erectus learned to use stones, spears, rope, and eventually fire. The tools allowed us to get more food; the fire allowed us to cook food that made us fatter; our brains used those calories to grow; and this facilitated the most revolutionary technological development of all—language. It’s harder for a woolly mammoth to kill you, Kislev reasons, when you can warn your uncle or your cousin that a mammoth is about to kill you. Language allows humans to work in groups, to hunt, cooperate, and love each other more deeply.

The dawn of farming and agrarian life around 12,000 BCE brought with it permanent settlements, and permanent settlements meant that humans no longer had to run around in small clans looking for nuts and berries, but could copulate inside of clay huts and have children who built clay huts nearby, thus giving us the “multigenerational family” (Relationships 2.0). Then steam power, sewing machines, chemicals, cement, and trains during the Industrial Revolution brought people from the country to the city, away from their uncles and cousins, and gave us the nuclear family (Relationships 3.0). And since the mid–20th century, we have been living in “the information society”: decentralized, multi-networked, hyper-individuated, global, multicultural, etc., etc., etc. (Relationships 4.0). Today, in the world of AI, quantum computing, extended reality (XR), and social robotics, we are stumbling into the future of Relationships 5.0: a “Super Smart Society” in which “robots will be available for an increasing range of bodily acts, ranging from simple housekeeping tasks to hugging and even sexual intercourse.”

This fingerpainting of human history puts Kislev’s project in league with those other “30,000-foot books,” as Anand Giridharadas once called them: the ones by writers like Yuval Noah Harari and Steven Pinker and Jared Diamond that are not unlike a man on an airplane who offers to enlighten you with an “explanation of everything.” The difference is that Kislev sprinkles in his original social-scientific data, expert interviews, and empirical research—which in some cases just involves his reading online product reviews—and so his project has a slightly meeker and more academic gloss.

What’s disconcerting is that Kislev doesn’t seem to register how depressing the world he’s describing is. For instance: Elon Musk’s company, Neuralink, which is developing a biotech implant that can read your mind, is seen by Kislev as a positive development. He rather cheerily imagines a scenario in which Neuralink might collaborate with a chatbot named Vincent after something traumatic happens to you, such as dropping an “irreplaceable china dinnerware set” on the ground. “You are simply inconsolable and totally irritable,” Kislev writes. “You do not know what you need to hear to feel better, so you are momentarily helpless.” Then the AI kicks in:

Without any verbal command or active input, the [Neuralink] chip could feed information about your personality, needs and desires to Vincent. In turn, Vincent will produce spontaneous text that is compatible with your exact needs and say the only thing in the world that could soothe you at this moment. Or, it can bring up a photo from a great trip you had and put on your favorite music, a song that, based on the data gathered, is very effective in soothing you in such cases.

It is hard to imagine a worse future for human beings: one in which a company run by Elon Musk not only gets direct access to your mind but systematically works to expunge all varieties of experience that might be considered challenging or uncomfortable. “What would have taken a trained psychotherapist or best friend months or years to learn how to comfort you,” Kislev writes, “Vincent figured out in a matter of milliseconds.” But what kind of infantilizing relationship is this—a friend spending years learning how to comfort you over a broken china set? It’s worth remembering that the aim of psychotherapy is not to make a patient happy but to enable them to experience unhappiness ordinarily.

One of the major casualties of the 5.0 world is language itself. In an interview that Kislev conducts with Guy de Beer, the cofounder of Kami Computing—an AI start-up that developed an “artificial persona capable of voice, long-term memory, advanced cognition, and emotions,” and that passed the legendary Turing test in 2019—de Beer reveals the “dirty secret” of the AI industry: that “the quality of conversations today is decreasing anyway, so the work of developers is easier.” Lexical patterns and styles of expression have become so flattened by texting and WhatsApp and auto-completed e-mails that it’s getting easier for robots to seem human. In other words, computers are not just getting smarter; they’re making people stupider. Somehow this is considered a sign of progress.

There is perhaps one vein of linguistic richness that’s come out of this new regime, and it’s the names of the apps and bots: Jarvis, Tug, Tomi, Zoosk, Zubee, Ying-Ying, Eugene Goostman. My personal favorite is Rollin’ Justin. He has opposable thumbs, can reach objects on high shelves, and can even make a cup of coffee—something that humans have been doing for about 600 years, give or take.

For a book supposedly about human-robot relations, one that starts with the story of a Chinese engineer marrying a robot named Ying-Ying in 2017, there is surprisingly little in Relationships 5.0 about sex. One of the exceptions is Kislev’s account of the company Realbotix, whose prototypes, Henry and Harmony, represent the frontier of sex robotics, or at least did in 2018. As Allison P. Davis reported in New York magazine:

Henry is outfitted in a white A-tank, sneakers, and Under Armour joggers that showcase his current penis attachment, which is 11 inches and nearly touches his knee. He’s about six feet tall in his stand, and he’s propped in the same posture as a caveman in a diorama at the Museum of Natural History. He has a six-pack, green eyes (slightly askew), full pink lips, and a slack jaw.

Kislev can’t find good empirical data on the performance of these sex robots, so he turns to customer reviews. His main source is a reviewer named Brick Dollbanger, who breaks Harmony while testing her. “I kind of knocked her senseless, mechanically. I mean, I didn’t really do anything to the AI, but gear-wise….” In the debate over sex robots, proponents see them as a way of eliminating sex work, supplementing sexless marriages, and promoting learning and exploration in a “safe” environment. Opponents see them as promoting and facilitating some of the worst social behavior imaginable: diminished empathy, a lack of interest in reciprocity, and objectification, if not sexual violence (and sociopathy). There is already a backlash to the backlash, with sex robot advocates denouncing the prejudice against their “robosexuality.” One doctor has even compared this kind of prejudice—rather unbelievably—to homophobia and transphobia.

The main problem with Kislev’s book is that it offers up technology as a solution for the very disaster that tech has helped to create. Are Internet-based chatbots, XR, neurotech implants, and sex dolls really the cure for loneliness and mental illness? It could be argued that the Internet is itself “the implacable engine of addiction, loneliness, false hopes, cruelty, psychosis, indebtedness, squandered life, the corrosion of memory, and social disintegration,” as Jonathan Crary recently argued in Scorched Earth: Beyond the Digital Age to a Post-Capitalist World. For all the promises of revolution and exploration and greater ease, it is more than plausible that the benefits of recent technologies have been outpaced by the havoc they’ve wrought on the environment and our sanity.

If we want to understand what’s going on in our technologically infused love lives, I’m not sure that sociology—or at least the strain that Kislev uses—is our best descriptive and analytical tool. For this book, Kislev conducted extensive surveys and interviews that ranged across thousands of people, and it’s unclear if any of them had something new or interesting to say. The responses he quotes are mostly civilian opinions about AI-human relationships that can be summarized in three words: pro, ambivalent, against. One gets the sense that we could learn more about the texture of what it’s like to love online by turning on a computer or reading tech journalism. Otherwise, we need approaches that are more creative or critically engaged—something that shows our world to us in a new light or suggests how we might reimagine it.

Relationships 5.0 depicts a world that is too familiar. It’s a place where people do not simply use computers as tools to love other people, but love the computers themselves at the expense of other people. Kislev assumes that this new society of Jarvis and Eugene Goostman and Ying-Ying is our “imminent reality”—that there is no alternative. But shouldn’t we aspire to something different? Many of us certainly want to. My favorite thing about Jarvis is that I don’t have to sit next to him on an airplane.

Dear reader,

I hope you enjoyed the article you just read. It’s just one of the many deeply reported and boundary-pushing stories we publish every day at The Nation. In a time of continued erosion of our fundamental rights and urgent global struggles for peace, independent journalism is now more vital than ever.

As a Nation reader, you are likely an engaged progressive who is passionate about bold ideas. I know I can count on you to help sustain our mission-driven journalism.

This month, we’re kicking off an ambitious Summer Fundraising Campaign with the goal of raising $15,000. With your support, we can continue to produce the hard-hitting journalism you rely on to cut through the noise of conservative, corporate media. Please, donate today.

A better world is out there—and we need your support to reach it.


Katrina vanden Heuvel
Editorial Director and Publisher, The Nation

Ad Policy