Love and Sex With Robots:

David Levy has spent much of the last 30 years thinking about how computers play chess. An international master of the game himself, Levy is the author of about 40 works on the topic, including Chess and Computers, The Joy of Computer Chess and All About Chess and Computers. In his latest book, Love and Sex With Robots, he turns his attention to another human activity machines can perform, and while love's wins and losses may be too fuzzy to allow a watershed computer victory comparable to Deep Blue's trouncing of chess champion Garry Kasparov in 1997, Levy believes their impact on the game of love will be equally triumphant. The author, it should be noted, is an outspoken advocate of artificial intelligence, and while the more primitive sexbots he imagines might be no more intelligent than Sony's robot dog, Aibo, he reserves his greatest hopes for truly intelligent artificial personalities. These would be capable of passing a "Turing Test," which means imitating human behavior so closely that we cannot tell we are interacting with a machine.

The book, based on the author's recent PhD thesis, presents evidence from a variety of psychological and technical fields in support of the central thesis that by around 2050, "love with robots will be as normal as love with other humans." Levy begins with a study of the reasons why humans fall in love. Scientific studies suggest we most often love people who possess certain characteristics such as similarity to ourselves, an attractive body and personality, the ability to fulfill our needs, physical proximity and perhaps a hint of mystery. These findings may seem like common sense, but Levy puts them to a surprising use. When we break love up into these component parts, he observes, we see that an artificially intelligent machine could be programmed to provide each one.

He does not shy away from the details of how this could be done. He bases a robot seduction strategy on a study by Rutgers University anthropologist Helen Fisher, who attempted to quantify the experience of love by conducting fMRI brain scans on students as they gazed at images of their partners. Fisher found increased activity in areas of the brain with high concentrations of receptors for dopamine, a chemical associated with euphoria and addiction. "A robot who wants to engender feelings of love from its human," Levy speculates, "might try all sorts of different strategies in an attempt to achieve this goal, such as suggesting a visit to the ballet, cooking the human's favorite food or making flattering statements about the human's haircut, then measuring the effect of each strategy by conducting an fMRI scan of the human's brain." The robot would know it was on the right track when it saw brain activity in the appropriate areas, and continue the successful strategy.

Would people really pay to have their most intimate emotions targeted in a manner reminiscent of a military drone searching for the weakest point in the enemy's defense? Levy makes a thorough case that the answer is "yes." He offers extensive evidence that we already form deep and lasting bonds with our pets, with our computers and televisions, and with people we know only from chat rooms on the Internet. Each of these examples illustrates that humans are not only capable of love outside the traditional setting of two human beings in physical contact, but enthusiastic about it.

But wouldn't people find love and sex with a machine lacking in emotional connection? Perhaps, Levy grants, but that probably won't stop us, just as it doesn't stop people from visiting prostitutes. Respondents to studies consistently report that they visit prostitutes as much for a feeling of emotional closeness as for pure sexual pleasure. These feelings are often illusory, but the johns and janes don't seem to mind. There's little reason to think this type of emotional projection, which Levy calls the "myth of mutuality," wouldn't carry over to robot partners as well.

Levy pulls data so exhaustive, and organizes them with logic so foolproof, that even the most skeptical will probably be convinced that robot sex is on the horizon. In fact, he verges on overkill, as many readers may not be so incredulous as he imagines. People living in rich countries today are already accustomed to socializing with and through machines. Young people, especially those under 20, conduct much of their social lives through Facebook and Myspace, interaction with artificial personalities — from robotic call-service workers to video game characters — is commonplace and Internet porn is hugely popular, generating somewhere between $366 million and $1 billion a year. We are well-primed for the transition Levy describes, and many readers will grow exasperated with his constant attacks on "the commonly held view that only sex with someone with whom one shares genuine affection can be a worthwhile and enjoyable experience."

The more interesting question is not whether people can accept sex with robots — we are nearly there already — but rather what its effects will look like and what they will mean for our future. Unfortunately, Levy answers this question like a robot-industry lobbyist, with no acknowledgement of potential problems outside the (serious) ethical issue of using artificially intelligent robots as sex slaves. His list of benefits includes improved sexual skills as robot sex coaches teach them to us, incredible orgasms, more successful marriages as robots sop up our spouses' excess desires and better understanding between the sexes. "I believe that the social and psychological benefits will be enormous," Levy proclaims. "Almost everyone wants someone to love, but many people have no one. If this natural human desire can be satisfied for everyone who is capable of loving, surely the world will be a much happier place."

It's easy to be swept along by Levy's enthusiasm, but such statements could benefit from a dash of history. Fifty years ago, for instance, many baby boomers were taught in schools that robots would make the toil of repetitive labor obsolete. Instead, the number of hours the average American works has stayed roughly constant since the end of the Second World War, while the employment-to-population ratio steadily increased from 52 percent in 1950 to about 60 percent in 2000. This means we're actually working more than we did before robots came on the scene, and the reasons are not technical but cultural: without any way of sharing the money saved by having robots do certain jobs, the promised benefits never arrive for most of us.

It's not difficult to imagine how our culture might interfere with Levy's breathless predictions as well. Some individuals will surely do as he predicts and use sexbots to improve their human relationships, but others may come to prefer a form of interaction that offers many of the benefits of love without any of its compromises and disappointments. Will people accustomed to the frictionless courtship of robot love find humans more bother than they're worth? In an age when people already seem to spend more time with televisions and computers than with one another, what does it mean to eliminate sex as an incentive for human engagement?

If Levy's unremitting optimism is troubling, even more so is his belief that robot mimicry of emotional behaviors is equivalent to human emotions. "The robot that gives the appearance, by its behavior, of having emotions should be regarded as having emotions," he writes, and again, "If a robot behaves as though it has feelings, can we reasonably argue that it does not?"

The author deserves credit for stating his position on this so directly, but it's not entirely clearly that he's thought out its full implications. Feelings exist only in our minds and we cannot directly experience those of another. However, as human beings we know that our desires emerge out of our history, our families, our politics, our imaginations, our pain and our sense of beauty. When we share our desire with another — even a prostitute or an avatar in an Internet chat room — we assume their desires run deep as well, even if they are only pretending to feel them for us. To say that the appearance of emotions should be treated the same as having them erases an element of free will crucial to the human experience.

But this equation also diminishes the robots, whose capacities need to be explored on their own terms. What if they open spaces in which human desires we've always been afraid of — sadistic, masochistic or other — can finally be enacted? What if their sexuality forces us to ask questions we’ve never asked before? And what if these questions stem not merely from the moments when the robot's artificiality is concealed, but from those moments when it becomes obvious as well? Imagine a lover who runs out of batteries in the middle of a sex act, for instance, or a lover whose eyes are as steady as surveillance cameras during sex. What will these experiences mean to us? What passions and paranoias will we discover through them?

Levy is right in his main points but his interpretations leave much to be desired. Barring a nuclear war or a full-scale economic collapse due to climate change, robot sex is very likely in the cards. But sex and love are not chess. They're not games people simply win or lose. In his desire to cry from the rooftops that the robots will bring "Great sex on tap for everyone, 24/7," Levy simplifies our desires and ignores the darker corners of a future that may just come to be.