Sex with robots, the moral and legal implications

As technology progresses, says Neil McArthur, there will be people immediately ready to exploit new forms of robotics in order to produce sexbots — robots explicitly designed to bring sexual satisfaction to humans. The University of Manitoba philosopher of applied ethics and social justice says it will raise concerns for some.

McArthur is the co-director of the U of M’s Centre for Professional and Applied Ethics, and writes frequently about the ethics of sexuality. He recently penned an article exploring the evolution of sexuality for British magazine Aeon, and he published an essay on Academia.edu — Guiltless Pleasures of the Lonely Human Being, The Moral and Legal Implications of Sex with Robots —   that explores the ethics of humans having sex with robots. You can find this essay and others on his blog moralust.ca, but we post an excerpt of the essay here.

The Moral and Legal Implications of Sex with Robots

Though they are a common feature of movies and television shows depicting the near future, human-like robots are currently at a very early stage of development. They await, in particular, two developments before they will begin to approximate the androids portrayed by Hollywood. First of all, they require advances in artificial intelligence, in order to produce interactions that provide a reasonable approximation of inter-human contact. Second, they require the development of synthetic materials that convincingly resemble flesh and mucosal membranes. However, a significant amount of research is underway in both areas, and major advances could come at any time. And we may conclude with certainty that, as technology progresses, there will be people immediately ready to exploit new forms of robotics in order to produce “sexbots” – robots explicitly designed to bring sexual satisfaction to humans.

For many, the prospect of sexbots raises concerns. The (meagre) public opinion data that is available suggests that the majority of people currently view the idea of sexbots with trepidation. I propose to examine the arguments for supporting or opposing the development of sexbots. I will argue that the overall effect of sexbots will, once they are widely available, be positive, and that we should encourage their development, promote positive attitudes towards them, and combat stigmas against their consumers. Nevertheless, the concerns about possible negative effects are not unreasonable, and they cannot, at this stage, be definitively dismissed. Instead, they require further research, and I believe an examination of the ethical arguments for and against sexbots can help guide such research. A discussion of sexbots is additionally valuable in that it allows us to explore questions about sexual morality more generally. People’s attitudes towards sexbots can be seen as reflecting a profound division in our attitudes towards sex, between those who view it as a bodily pleasure like any other, and those who believe it should have special significance.

Some groundwork: Robots and the Significance of Sex

In February of 2013, the polling firm YouGov conducted a poll, sponsored by the Huffington Post website, that asked people about their attitudes towards robots. It was a relatively large (1,000 adults), random sample. Asked if they would want a robot caring for an ageing friend or relative, fifty-eight percent said no. More relevant to our purposes, another question asked: “If it were possible, would you ever have sex with a robot?” Only nine percent said yes. Eleven percent said they were not sure, and the remaining eighty-one percent (figures were rounded by the polling firm) said no. The poll also asked: “if it were possible for humans to have sex with robots, do you think that a person in an exclusive relationship who had sex with a robot would be cheating?” 42% said yes, and 26% said they were not sure. Only 31% said no. The polling data unfortunately does not give us any specific insights into why people answered the way they did. I do not intend to speculate about the reasons people may, as a matter of fact, have for their views. Rather, I would like to explore how different views about sexbots might be grounded philosophically.

In order to do so, I would first like to lay out a fundamental division in people’s views towards sex more generally. David Benatar has distinguished between two views of sexual activity, which he calls the Significance View and the Casual View. According to the first of these: “for sex to be morally acceptable, it must be an expression of (romantic) love. It must, in other words, signify feelings of affection that are commensurate with the intimacy of the sexual activity. On this view a sexual union can be acceptable only if it reflects the reciprocal love and affection of the parties to that union.”

According to the Casual View, by contrast: “Sexual pleasure . . . is morally like any other pleasure and may be enjoyed subject only to the usual sorts of moral constraints.” I think Benatar is uncharitable to label this second view as “casual”, since that term implies that those who subscribe to it think that sex is never, or should never be, significant. In fact, few if any people who hold Benatar’s “casual view” would deny that sex may have special significance as part of a loving relationship. They merely reject the proposition that significance should be a necessary condition of moral acceptability. A well-known proponent of this view, Logan Levkoff, asserts: “I do not believe that anyone (man or woman) has to engage in a relationship in order to have sex of any kind. Everyone is entitled to have emotionally unencumbered hookups.” The term casual is inaccurate for the further reason that people who hold the view often attach great importance to sex. People who hold this view of sex often refer to themselves as “sex positive”, and I will accordingly refer to it as the Positive View. For convenience (and at the cost of elegance), I will call those who subscribe to the first view Significants, and those who subscribe to the second, Positives.

To defend their view, Positives appeal to a principle they take to be generally accepted, and indeed at the basis of a democratic society: the harm principle. According to this principle, which was most famously formulated by John Stuart Mill, people generally have the right to live their lives as they see fit, and to pursue their own conception of the good. We are only entitled to condemn a person if we can show that her behaviour causes demonstrable harm to someone else. This principle entails a moral right to privacy. Where our actions do not affect others, they are not properly the subject of moral judgement.

Significants do not ignore the value of personal autonomy, which provides the basis for the harm principle. They simply reject it as the exclusive basis for moral judgement. It is easy for Positives to misunderstand the Significance View, and thus miss the nature of the disagreement between the two views. Many Positives simply take for granted that sexual behaviour, when it concerns two consenting adults, is not properly the subject of moral judgement at all. As far as they are concerned, Significants are free to restrict their own sexual relations to loving relationships, but the rest of us should be equally free not to. However, Significants cannot accept this. According to my definitions, if someone restricts her own sexual activity to significant sex, but she is willing to accept that there is nothing immoral about others having non-significant sex, she is in fact a Positive. As I have defined the two positions, Significants (properly understood) believe that non-significant sex constitutes a moral wrong even when the participants are both freely-consenting adults – and even when neither is a Significant….

All those in favour say “Aye, Robot”

Positives and the Case for Sexbots

From the perspective of the ethics of autonomy, which is adopted by Positives, it is hard to see anything wrong with sex with a robot. According to the harm principle, if people want sexbots and they do no measurable harm, we have no grounds for judging people who use them. And more than this, Positives can give reasons to see the development of sexbots as a positive good. First of all, sexbots promise to deliver direct hedonic benefits. All other things being equal, sex is considered by most people to be a good rather than a bad thing. A life with more sex is generally preferable to one with less. And various studies have suggested that people generally get less sex than they would like, and would be happier if they had more. One study has concluded that for the average person, increasing the frequency of sexual intercourse from once a month to at least once a week offered as much additional happiness as an increase in salary of $50,000 per year.  It is reasonable to conclude that the possession of a realistic sexbot will, at least for many people, lead to an increase in the absolute quantity of sexual experiences. Sexbots thus have the potential to maximise both the amount of hedonic satisfaction, and as a result the level of overall happiness, in the world.

Greater levels of sexual satisfaction, apart from their direct hedonic value, contribute to better health outcomes more generally. High levels of sexual activity correlate to weight loss, lower stress levels, better heart and blood-pressure outcomes, lower rates of prostate cancer for men, and better sleep. People who have more sex quite simply tend to live longer, healthier lives. Some of these benefits can be achieved through solitary sex or the use of existing sexual aids. However, others are the result of the physical exertion required for sex with a partner, and sexbots could elicit comparable levels of exertion. Indeed, it is easy to imagine that the robots could be programmed specifically to provide their users with a work-out, at variable levels of strenuousness, at the same time as they provide sexual satisfaction. Considering that there are many people who want to keep fit but have trouble motivating themselves to do more conventional forms of exercise, sexbots could provoke a general improvement in people’s fitness. Research has shown that sex with a partner has certain psychological benefits that masturbation cannot achieve. It may be that many of these benefits are also achievable with robots. At the very least, this presents a question for further research. And even when it comes to those benefits that can also be achieved through masturbation, many people will, we may presume, find sex with a robot more satisfying than masturbation, which means it will be done more frequently, and the level of benefit achieved will thus be greater.

There are also egalitarian reasons for favouring the development of sexbots. We should note that not only are humans generally under-sexed, sexual satisfaction is unevenly distributed. There is more of it available to those who are good-looking, wealthy, intelligent, famous, or who have the good fortune to live in social environments where sex is more easily available, such as university dormitories and other areas with concentrations of single young people. At the other end of this scale are those who are mentally or physically disabled, or who are forced to live in single-sex environments such as prisons or mining camps. Some societies, most notably China, possess dramatically uneven sex ratios overall, which leads to large numbers of men with no opportunity for sexual companionship. Sexbots have at least the potential to equalise access to sexual fulfillment, ensuring that everyone who wants it can have it more easily regardless of their natural endowments or social position. It might be argued that they will be quite expensive to own, which is surely true. But short-term access – in essence, “robot sex workers” – could potentially be quite affordable, and could even be subsidised if their benefits were widely acknowledged.

Sex with robots will not, in the short term, deliver the same levels of hedonic satisfaction to most people that sex with a human partner does. But they are better than nothing, and as the technology improves, the hedonic gap between robot sex and “normal” sex will close. If we are Positives, we thus have good reason to view the development of sexbots as beneficial, and worthy of encouragement.

Turn (it) off

Significants and the Case against Sexbots

Let’s keep it platonic.

Significants may have several reasons for finding robot sex objectionable. First of all, such people might say that robot sex, as an act of mere sexual release, is inherently degrading. A Positive might ask how robot sex is different from mere masturbation. However, for many Significants this may be precisely the point – sexbots serve as an incitement to masturbation, and they consider masturbation immoral. Kant was a Significant, for instance, and he condemns masturbation as an act in which a person yields herself over to her animal nature and uses her body merely as an object of pleasure. As he puts it: “onania . . . is a misuse of the sexual faculty without any object. . . . man thereby forfeits his person, and degrades himself lower than a beast.”

This view of masturbation is held by modern philosophers as well. For Roger Scruton, masturbation is immoral because it “involves a concentration on the body and its curious pleasures” – indeed, an “obsession . . . with the organs themselves and with the pleasures of sensation.” John Finnis also holds the Kantian view (which he draws from Aquinas and natural law theory) that masturbation entails a surrender to one’s purely “physical self”, and thus an abandonment of the “choosing self” that makes us human. It is, for this reason, a degradation of our nature. If one holds this view of masturbation as degrading, then precisely those facts about sexbots that, for Positives, make their development a moral good, should give us cause for concern. Because they incite and indeed amplify the pleasure to be had from mere masturbation, sexbots will increase the aggregate levels of auto-eroticism in society, and intensify the surrender to one’s bodily pleasures that masturbation entails.

Not all Significants object to masturbation. However, even those Significants who view onanism as morally neutral may have reasons for objecting to sexbots. One is that it is not just solo sex, it is sex with a machine. It could be argued that sex with a machine, like sex with food or sex with animals, is an affront to the special dignity to which we should properly attach the human sexual act. One might consider the public reaction to the movie “American Pie”, the title of which comes from a scene which a teen-age male has intercourse with a pie, to see that sex with non-human objects fills many people with a disgust that mere masturbation does not. When I discuss sex with robots with people, they often wonder how a sexbot is different than any other sex toy. But for some Significants, this may be precisely the point. We have only to look at the several U.S. states that have tried to ban sex toys – which the laws revealingly refer to as “obscene devices” – to see that many people consider it inherently objectionable to use machines for auto-erotic purposes.

I have said that many Significants also hold a teleological view of sex, according to which sex is morally acceptable if it serves a particular end or ends. There are three chief ends that sex may serve: (a) reproduction, (b) the building of relationships, and (c) through its role in supporting marriage, the strengthening of communities. Significants may condemn robot sex for failing to serve, or for working contrary to, these ends. Even those Significants who are tolerant of masturbation – and even those who do not object to sex toys – may have reason for objecting to sexbots on teleological grounds.

A. Reproductive Ends

Many people see sex, to the extent that it can be morally sanctioned, as being essentially linked to reproduction. It is, as John Finnis, puts it, “an instrumental good” that is meant to be “in the service” of procreation. Such people do not generally insist that each individual sex act must be intended to produce children. Rather, they argue that acceptable sex acts must belong to the class of acts that could potentially be reproductive. Kant says that “natural sex” is that as a result of which “procreation of a being of the same kind is possible”. Using the language of natural law theory, Finnis says that sex is morally acceptable to the extent that it instantiates a “biological union” between a man and a woman. He defines such a union in this way: “Biological union between humans is the inseminatory union of male genital organ with female genital organ; in most circumstances it does not result in generation, but it is the behaviour that unites biologically because it is the behaviour which, as behaviour, is suitable for generation.”

Robot sex obviously cannot be reproductive. Significants may see it as being on par with gay sex, to which many Significants object more strongly than masturbation. Gay sex is “real” dyadic sex, and therefore belongs to the general class of acts that should instantiate a biological union. Yet it cannot in fact instantiate such a union. Significants may see sex with robot in the same way – as sufficiently similar to dyadic heterosexual unions to belong to the general class of “real” sex acts, without allowing for any possibility of biological union.

We might raise an objection here: why, we might wonder, does robot sex count as “real”, dyadic sex, rather than as mere physical release? One answer is, because people will tend to see it as such, because of the way in which they tend to view robots. We can look here at research on research by Sherry Turkle, who has argued that robots are different than other forms of technology because they are what she calls “relational artifacts”. Relational artifacts are non-living objects that are, or at least appear to be, sufficiently responsive that people naturally conceive themselves to be in a mutual relationship with them. They therefore affect our sense of self, our emotional well-being, and our relationships with other humans, in ways that other forms of technology do not. Turkle has looked at studies of people interacting with robots in various settings, and this empirical work suggests that many of us have a natural, even irresistible, tendency to project human intentions and emotions onto such objects, to attribute to them some sort of soul or essence, and to form an emotional bond with them. She points to one study of human-robot interaction that observed a high percentage of their subjects attributing to robots some kind of  “technological essence” (75% made this attribution), a “life-like essence” (48%), mental states (60%), and social rapport (59%).

If it is empirically true that people tend to see robots as a special kind of object, sex with a robot potentially becomes more than a mere auto-erotic act. It includes an entity with which we naturally, at least in our own minds, form a certain kind of relationship, and which is therefore distinct in kind from, and more dangerous than, other sorts of sex toys. We are having “real” sex with an entity with which we can never form a true, reciprocal bond.

B. Relationship Ends

Significants identify another end that makes sex within a loving relationship morally acceptable: the building and strengthening of the relationship itself. Says one writer: “Both human experience and biology make it clear that sexual intercourse, which is essential for the consummation of a marriage, has a twofold telos: strengthening the bond between partners and creating children.” Finnis says that marriage brings a “double blessing–procreation and friendship—[that] is a real common good . . . that can be both actualized and experienced in the orgasmic union of the reproductive organs of a man and a woman united in commitment to that good.”

Sex with a robot cannot create such a union. Robot sex might even work against the purpose of building intimacy between partners, either by displacing the desire we have for our partner or by discouraging us from meeting a partner in the first place. Robot sex may be seen as particularly dangerous from this perspective, much more so than mere masturbation or the use of sex toys, because, as Turkle’s research shows, it may offer a dangerous illusion that an intimate union is being achieved. Significants might worry that in succumbing to such an illusion, we are harming our ability to form actual, loving bonds with other human beings. Though she does not discuss sex with robots directly, Turkle has raised the alarm about the possible impact of robots companions. She says that in making such robots:

in the process, we are remaking human values and human connections. We change ourselves, even before we make the robot. We think we are making robots, but we are remaking people . . . . When we assume artificial companionship, it changes how our children grow up, it changes how we treat each other, it changes how we think about caring for each other, across the generations.

She concludes: “Dependence on a robot presents itself as risk free. But when one becomes accustomed to ‘companionship’ without demands, life with people may seem overwhelming.” If Turkle is right, sexbots may potentially make it more difficult for people to form relationships. And for people in relationships, the bond they form with their sexbot – uncomplicated and risk-free as it seems – will compete with the messier, more fragile bond they have with their own partner, often to the detriment of the latter.

C. Community Ends

If sexbots do undermine people’s capacity to form loving relationships, and thus to start families, they also work against another, related end that some Significants think sex should serve. For communitarians – people who prioritise the importance of close, integrated communities, and the sense of collective identity that membership in such communities brings with it – the loving relationships that sex helps to build, and the families that result, are at the heart of a well-functioning society. Many communitarians see the family unit as the primary vehicle to transmit social values. As Scruton puts it:

Marriage is not merely a tie between man and woman; it is the principal forum in which social capital is passed on. By tying sexual fulfillment to the bearing of children, marriage offers a double guarantee of a stable home: the guarantee that comes from erotic love, and the guarantee that comes from the shared love of offspring. It offers children durable affection, a secure territory, moral examples, and moral discipline . . . All of us therefore have a deep and lasting interest in marriage, as the only known way to reproduce the moral order.

If sexbots lead fewer people to form families, and if they de-stabilise those families that are formed, the social fabric will be weakened as a result.

Conclusion

I have not discussed the question of the moral status and legal rights of robots themselves. The question of moral and legal personhood for robots and other possessors of artificial intelligence has already received significant attention in the scholarly literature. To enter into this debate would go beyond the scope of this paper. I will merely say that sexbots will be developed that meet the needs of their owners in terms of personality and intelligence long before they reach the point where most people would consider them legitimate candidates for personhood. However, if robots develop to the point that we think they might possess human-like minds, the thought that they are being kept in large numbers as sexual servants may be particular disturbing to many people, and will bring to the fore the general question of robots’ moral status.

Sexbots are on their way, and, since their development will proceed in tandem with more general advances in robotics, it would be virtually impossible to prevent this development even if we wanted to. However, our assessment of their moral value will help determine whether they become an accepted part of our culture or, instead, an underground technology that people view with shame and embarrassment. I have argued that we should embrace, and encourage, their development. Given current attitudes, early adopters of sexbots may face stigmatisation – as indeed may be those involved in their development, production and distribution. We should oppose such stigmatisation. I have argued that “common ground” concerns about sexbots depend on empirical propositions that are either unproven or insufficient to outweigh the positive arguments. However, given the uncertain state of the evidence in most of these cases, and the fact that we can only look at evidence for analogous phenomena, none of the concerns can be ruled out. We must be prepared to revise our views based on new data. And indeed, we require more data. An understanding of possible ethical concerns about sexbots can help us better direct future research.

Neil McArthur

Neil McArthur is an associate professor in the department of philosophy, and the associate director of the Centre for Professional and Applied Ethics at the University of Manitoba. The above has been edited for length. To read it in full, with all academic citations, please click here.