Love and Sex with Robots_The Evolution of Human-Robot Relationships

by David Levy

  0.69

  To have children

  0.49

  TABLE 3

  The majority of the motivations listed in both of the above surveys are presented in somewhat egocentric terms, generally indicating something that the respondents want for themselves out of the sexual experience. In contrast, a third study, in 2004, which was based on a survey conducted via the Internet, expressed ten of its eleven proffered motivations in terms of the way the respondents related to and felt about their partners. These were the following: to achieve emotional closeness, raising one’s self-esteem by increasing the feeling of being desirable and wanted, to nurture and care for the partner, experiencing the partner’s power, to obtain approval and reassurance from the partner, to disarm the partner and protect oneself against hostility or the partner’s negative moods, to exert power or control over the partner, and to elicit nurturing and caregiving from the partner.

  TABLE 4

  In all three of these studies, pleasure and emotional closeness were at or very near the top of the respondents’ lists, so we shall examine them first when considering why humans might want to have sex with robots. Let us start with pleasure.

  The most obvious way in which humans obtain pleasure from sex is through orgasm, and a robot that can give its partner great orgasms on demand will therefore be highly prized as a sexual partner. In chapter 7 we shall discuss the technologies that will most likely be employed in robots for achieving this goal, but in the meantime consider how a simple sex doll, with no electronic brain, no artificial intelligence, and none of the humanlike characteristics that come from these technologies, can help men to achieve great orgasms. In 1997 the popular radio “shock jock” Howard Stern was given the “Celine” model of a RealDoll by its manufacturer, Abyss Creations. Stern tried it out and waxed lyrical on his radio show about the experience, proclaiming it to be “the best sex I ever had! I swear to God!”

  This primitive example shows how much pleasure a mere doll can bring its owner, through the simple expedient of being a nonactive partner in an orgasmic experience. Despite Stern’s claim that this was the best sex he’d ever had, just imagine how much better it might have been for him if instead of a lifeless doll, with no intelligence, no conversation, and no sparkle, he had pleasured himself with a fembot who told him how much she loved him and what a wonderful lover he was, who caressed him and employed her other sensual capabilities to heighten his enjoyment of the encounter. But more about sex dolls and the technology of the orgasm in chapter 7. For now we return to consider the other reasons, apart from the pure physical pleasure of orgasm, why we enjoy sex.

  I feel certain that some readers, despite having digested the evidence and arguments in the earlier chapters, do not yet believe that rational humans will develop emotional attachments for robots by midcentury, let alone be falling in love with them. But those readers will surely admit that a fembot or malebot who not only gives great orgasms but also relieves one’s sexual tensions, provides new sexual experiences, leads a path away from boredom, and reduces stress could make an outstanding lover. So even in the absence of a strong emotional attachment from the human side, there will be ample motivation for a significant proportion of the population to desire sex with their robots. For example, the 60 percent of college men in James Carroll’s survey who did not respond that for them emotional involvement was always or most of the time a prerequisite for having sex—they will be likely customers. Similarly will the 51 percent of men in DeLamater’s study who mentioned lust and physical pleasure as their main motivations for engaging in their most recent sexual encounter.

  Those readers who do already accept the concept of humans falling in love with robots can add to the list of benefits in the previous paragraph more of those in Tables 2 to 4—the ones derived from the emotional attachment that loving owners will feel for their electronic sex partners: the expression by the robot’s owner of their closeness and/or love for the fembot or malebot, the giving of pleasure to the robot partner, obtaining reassurance about the robot’s virtual love for its owner, the enhancement of one’s self-esteem on being praised by the robot for one’s lovemaking skills, and satisfying the robot partner’s stated desire for sex.

  Even for those motivations found in the lower reaches of the survey statistics—those relating to the human’s power and domination of the sex partner, giving the human a feeling of sexual conquest, and to drink or drugs as the stimulus for having sex—there will certainly be some occasions when these motivations provide sufficient impetus for having sex with a robot, even for people who are not impelled by any of the more powerful motivations.

  This leaves only one motivation from Tables 2, 3, and 4 that cannot be applied to human-robot sexual activity—the desire to procreate with a robot.*

  Thus far in this chapter, we have explored the main reasons people decide to have sex with people and why people in the decades to come will decide to have sex with robots, but this discussion has somewhat ignored an important catalytic effect that increases the likelihood that a sexual encounter will take place—the sex appeal of the prospective sex object. In this sense the bare statistics expressed by the survey respondents present only part of the picture. The other part, the seductive part, is less obviously a reason to a participant in a psychological survey, largely because of the natural inclination to rationalize when answering a questionnaire rather than to admit to being influenced by factors that are not directly related to sexual decision making.

  These “other factors”—the behavioral ones, the seductive ones—have been investigated by David Bass at the University of Michigan, who ranked various “male acts” and “female acts” according to how effective they were assessed to be in leading the person’s date to the bedroom. In Bass’s list of the “twenty most effective male acts,” there are nine that could apply to robots, including the top three.

  1st—He displayed a good sense of humor.

  2nd—He was sympathetic to her troubles.

  3rd—He showed good manners.

  6th—He offered to help her.

  14th—He smiled a lot at women.

  15th—He gave encouraging glances to girls.

  18th—He touched her.

  19th—He made up jokes to make women laugh.

  20th—He expressed strong opinions.

  In any of these assessments, gleaned mostly from the comments of close friends of the women who admitted being influenced by these “acts,” we can replace “he” (the woman’s sexual partner) by “it” (a robot) with no loss of validity. Already there are computer programs that can make up new jokes.* Most of them are not wonderful jokes, but some are clever enough to get a smile or a laugh. As the software technologies of joke making develop, so robots will come to appreciate jokes made by their conversation partners—in other words, to have a sense of humor. These robots will ably perform the first and nineteenth acts on the above list.

  Being sympathetic, well mannered, and helpful and being able to express strong opinions during conversation are attributes that come from a combination of empathy and conversational skills—nothing here is beyond the bounds of reasonable expectation for the artificially intelligent robots of 2050. As for smiling and giving encouraging glances, David Hanson’s moving head* can already accomplish both. And touching, of course, is just about the easiest thing to design into a robot.

  The “twenty most effective female acts” listed by Bass strongly mirrored the male list, omitting only two acts—touching and expressing strong opinions—and including one act not on the male list—telling him things he wanted to hear (another straightforward task, once robots have reached the necessary level in their conversational skills).

  Sex as a Result of Transference

  Transference is a psychological phenomenon, typically described as a subconscious redirection of feelings from one person to another. Whereas attachment is a transference of positive feelings that develop first and specifically for a baby’s primary carer, u
sually its mother, and later in life to objects important to that baby/child/adult, and possibly to other people in the form of romantic love, transference is a redirection of feelings, positive or negative, that were first associated with a significant person in the subject’s life, not necessarily its primary carer, and are later transferred toward some other person. As an example, one might have negative feelings toward somebody whose manners, voice, or appearance resembles that of an abusive parent, a sadistic teacher, or physical education instructor, a bully or a tease, or a loathed exspouse. Examples of positive feelings also abound and might be more closely related to sexuality: a dazzling girl who sat in front of a boy in their high-school algebra class, inspiring his sexual fantasies, or a sexy teacher whose slit skirt and abundant cleavage were similarly inspirational. Transference was first described by Sigmund Freud, who recognized that the models we create in our minds during our formative years as to how people behave stay with us and affect our choices, experiences, and relationships into adulthood.

  John Suler, in his fascinating article “Mom, Dad, Computer (Transference Reactions to Computers),” published on the Psychology of Cyberspace Web site, explains how the phenomenon of transference extends to relationships with computers:

  These models also shape how people select and experience things in their lives that are NOT human, but so closely touch our needs and emotions that we want to imbue them with human characteristics. We humans can’t help but anthropomorphize the elements in the world around us. It’s in our blood. We use our internal models to humanize and shape our experience of cars, houses, pets, careers, the weather…and COMPUTERS.

  Yes, computers can be a prime target for transference, because they may be perceived as humanlike. They are complex machines that almost seem to “think” like humans think. In fact, some people say they WILL someday be able to “think” like us. Unlike TV, movies, or books, they are highly interactive. We ask them to do something and they do it—at least they usually do (like humans, they sometimes disobey and surprise us). With the new generation of highly visual, auditory, and customisable operating systems and software applications, we also have a machine that can be tailored to reflect what we expect in a companion. The science-fiction fascination with robots and androids is the culmination of this perception of machines as being almost like one of us.

  What makes computers especially enticing targets for transference is that they are VAGUELY human and PROGRAMMABLE to be whatever we make them out to be. Psychoanalysts discovered that if they remain relatively ambiguous and neutral in how they behaved with their clients, the clients would begin to shape their perceptions of the analyst according to their internal models from childhood. When faced with an indistinct, seemingly malleable “other,” we instinctively fall back on our familiar mental theories about relationships and use those theories to shape how we think, feel, and react to this new, somewhat unclear relationship. This whole process is often unconscious. We are so used to these old templates that they automatically start to mould our perceptions and actions without our really thinking about it.1

  Suler’s article continues with a discussion of the various ways in which transference can apply to computers—how we might subconsciously experience our computer as being like our mother or father or a sibling. One of these ways is examined in the context of Freudian psychology, in terms of sexual desires and fantasies experienced in relation to one’s parents, a subject explored more fully by Norman Holland in another article from the Psychology of Cyberspace Web site. Quoting Joseph Weizenbaum’s reaction in 1976 to the way that people anthropomorphized and became deeply involved with his programs ELIZA and DOCTOR, Holland points out that people form bonds with computers more quickly than with other objects:

  The computer just makes this process faster and more drastic, because it exhibits “intelligent” behavior like another human.

  In sum, then, we have some fantasies about the computer as a thing: phallic fantasies of power and oral fantasies of engulfing pleasure. We also have these more remarkable fantasies that the computer is something more than a thing, something between person and thing. We have a quasi-human relationship with the machine as helpmate, as true friend, as permissive parent, as sex object, and as sex partner.2

  Suler describes Holland’s view of the computer as being “seen as seductive, as a sex object, a satisfier of desire, as a symbol of sexual power and prowess.” Thus the concept of transference to computers has rapidly become a discussion of computers as sex objects, complementing our analysis of the reasons for having sex (with people) and the inference that the same reasons mostly apply to having sex with robots. This analysis demonstrates that humans have the capacity for desiring sex with a robot, the capacity to be seduced by them, while Suler and Holland show that computers, and hence robots, have the capacity to entice us, to seduce us.

  6 Why People Pay for Sex

  I pay for sex because it is the only way I can get sex. I am not ashamed of paying for sex. I pay for food. I pay for clothing. I pay for shelter. Why should I not also pay for sex? Paying for sex does not diminish the pleasure I derive from it. I am proud that I can afford to pay for as much sex as I need. Indeed, I sometimes pay more than asked or expected.

  —Hugh Loebner1

  The idea of satisfactory sex being available whenever it is desired is an extremely appealing one to enormous numbers of men and women, but for various reasons many people do not enjoy this level of sexual availability within the confines of a relationship or marriage. This might be because they are not in a relationship at all, or because they are often away from their partner because of business or other travel commitments, or because their relationship partner does not enjoy sex (with them) as much as they do (with their partner) or is not a good sex partner in some other respect. Whatever the reason, this void in their life has a simple remedy that for thousands of years has been adopted by a small but significant proportion of the sexually active population. The remedy is to pay for it, and the prevalence with which paying for sex has existed for so long has enabled the world’s oldest profession to survive and even at times to thrive. Soliciting the services of prostitutes provides a relatively easy cure for sexual frustration, and there is no evidence to suggest that those with the necessary financial means need ever go without sex through a lack of supply. Far from it.

  In this chapter we examine the reasons people pay for sex with the women and men who ply this particular trade. From the perspective of sex with robots, what is interesting about the most frequently proffered reasons is that they indicate desires, and not only the desires for the sex acts themselves, that could be satisfied by a sophisticated robot just as well as by a human prostitute. This being the case, it seems inevitable that just as humans desirous of sex but lacking sufficient opportunity will pay a professional for it, so there will come a time—and that time is almost with us—when people will be paying for sex with robots, either by buying the robot for regular use at home or by renting one by the hour or the day. The enjoyment and benefits derived by their owners or renters from the sex they experience with robots can reasonably be expected to bring as much overall satisfaction as those same people enjoy as the clients of (human) prostitutes.

  We commence our discussion of the parallels between paying human prostitutes and purchasing robot sex with a comparison between the reasons men pay for sex and the reasons women do so. The purpose of this comparison, in part, is to support the argument that women will become just as eager as men to seek sexual satisfaction from robots.

  Men Paying Women

  Obtaining accurate estimates of the percentage of the population that visits prostitutes is fraught with difficulties, largely due to the stigmatizing view of prostitution and its clients that has long been held by so many people. As a result, serious attempts at quantifying the use made of the services of prostitutes did not begin until the middle of the twentieth century, when Alfred Kinsey estimated that 69 percent of the white male population of Amer
ica had been to a prostitute at least once in their lives. If this figure seems high to some readers, it should be considered alongside the historic study by Timothy Gilfoyle, who estimated that 10 to 25 percent of all young women in New York City in the nineteenth century were prostitutes, either temporarily or on a long-term basis, making prostitution for much of that period the second-largest business in terms of the revenue generated (the first being tailoring). It should also be noted that Kinsey’s figure of 69 percent was for men who have had sex with a prostitute at least once in their lives, and most of these men had only one or two such experiences. Kinsey’s estimate of the proportion of American men whose sole sexual outlet was prostitutes was very much lower, between 3.5 and 4 percent. Other, more recent estimates by different researchers of the numbers of men in the United States who had had sex for money have ranged from 16 percent (estimated in 1994), to 18 percent of those aged eighteen to fifty-nine (in 1998), to 20 percent (in 1993).

  The accuracy of all these figures must be viewed with some doubt because of the known discrepancies between what johns are willing to admit in an interview survey or when filling in a pencil-and-paper questionnaire and the figures ascertained by other means that are known to be more reliable. This phenomenon of underreporting by johns was examined in a study published in 2000 by the U.S. Academy of Sciences, entitled Prostitution and the Sex Discrepancy in Reported Number of Sexual Partners. In their report, Devon Brewer and his colleagues found that when responding via a computer-assisted interviewing process, which is generally believed to promote accurate reporting, the johns’ answers relating to contacts with prostitutes were almost four times higher than when responding to human interviewing or pencil-and-paper questionnaires. It seems most unlikely (not to say impossible) that this four-to-one ratio persists across the whole spectrum of quantitative research on prostitution, but what is clear is that figures such as 16 to 20 percent in the United States should certainly be regarded as understatements.