Skip to content

Oxford Uehiro Prize in Practical Ethics: Is Sex With Robots Rape? Written by Romy Eskens

This essay was the winner in the Graduate Category of the Oxford Uehiro Prize in Practical Ethics 2017

Written by University of Oxford student, Romy Eskens

On The Permissibility of Consentless Sex With Robots

Recent movies and TV-series, such as Ex Machina and Westworld, have sparked popular interest in sex robots, which are embodied AI systems designed to provide sex for humans. Although for many it may seem absurd to think that humans will ever replace their human bedpartners with artificial machines, the first sexbots have already entered the commercial market. In 2010, TrueCompanion introduced Roxxxy, a sexbot with synthetic skin and an AI system that allows her to interact with her user through speech and affective communication. Another example of sexbots currently for sale are the RealDolls, which are silicone sexbots available in different models and upgradable with insertable faces and body parts. The question I address in this essay is: do humans require consent from sexbots for sexual activity to be permissible?

There are convincing ethical reasons to create sexbots. To begin with, sexbots can replace human sex workers, thereby reducing harmful practices such as sex slavery and sexual abuse.[i] Moreover, they can provide satisfying alternatives for individuals with sexual desires that could harm human beings if brought into practice, such as the desire to have sex with children or to engage in extremely violent or degrading sex. Furthermore, sexbots can provide a solution for individuals who experience difficulty in finding sexual partners, and can provide intimate companionship for those who feel lonely or isolated. Finally, sexbots can satisfy the sexual needs of convicts or soldiers, which is likely to decrease the instances of sexual violence in prison and in the army.

Although Roxxxy and the Real Dolls are still a far cry from the highly sophisticated robots featured in Ex Machina and Westworld, the current acceleration of AI development suggests that it lies within the scope of near future technical possibilities to create interactive sexbots with flesh-like skin, affective computing, highly developed sensory perception, refined language skills, the capacity to learn, and a pre-programmed personality. Such robots will be able to satisfy the sexual needs of their users, as they adapt to the sexual preferences of their users, and base their sexual performance on an amount of data never to be achieved in a single human life. Although sexbots will be programmed to act as if they have qualitative experiences, such as the experience of pleasure when having sex, they will in fact not be sentient. Still, this behavioural characteristic will make it possible for humans to engage in intimate relationships with them, and it is this quality that radically differentiates sexbots from sex toys.

Since sexbots will operate in an open environment, they must enjoy a degree of autonomy. However, it will not be technically possible in the near future, and many experts doubt that it ever will be, to incorporate the capacity for morally autonomous decision-making into the design of robots. For an agent to exercise moral autonomy is for her to act on rules she has imposed on herself, to which end sophisticated intellectual abilities, such as consciousness, reason-responsiveness and future-orientedness, are required. Sexbots will lack these abilities, and their capacity for autonomy will be restricted to decision-making within the action-directing parameters implemented by a programmer.

Since sexbots are thus bound to act within the parameters set by programmers, and they are programmed to provide sex for humans, they cannot choose to have sex. This means that they cannot agree with, or consent to, sexual relationships with their users. But there is a general agreement in both common sense morality and ethical theory that sexual activity that takes place without consent from one of the parties involved is morally wrong, which is also reflected in the juridical characterization of rape as sex without consent. Hence, it seems to follow that sexual relationships between humans and sexbots are impermissible, and that robots are wronged by carrying out the acts they are designed to perform.

Broadly stated, someone who agrees to have sex, either explicitly by speech, or implicitly by commencing sexual activity or performing a sexual act, consents to the act. On the basis of this definition, sexual activity appears wrong if it is forced on an individual who withheld consent, or if an individual has neither explicitly nor implicitly given consent to having sex, such as is the case when someone is unconscious. However, this broad definition leads to the implausible conclusion that cases in which an individual agrees to have sex under pressure, or is coerced to perform a sexual act, are cases in which consent has been given. In addition, incompetent decision-makers, such as persons who are under high influence of alcohol or drugs, as well as children and severely cognitively impaired individuals, sometimes agree to have sex. Yet these individuals lack the capacities for moral autonomy and rational deliberation, which are necessary for acquiring an understanding of the relevant facts, implications and consequences of this act, on the basis of which they can choose to withhold or give consent. Therefore, a narrower definition of sex without consent, defined as sexual activity that includes at least one individual who has not expressed informed and voluntary consent, better fits our moral judgments.

Now, I imagine few are willing to accept the conclusion that it is morally wrong to engage in sexual relationships with robots, whereas most do think that sexual activity is wrong if consent from a human being is lacking. In order to justify this intuition, it must be demonstrated that there is some morally significant difference between consentless sex between humans on the one hand, and consentless sex between humans and robots on the other hand.

One way to do this is to point out that the wrongness of consentless sex depends on membership of the human species. But to ascribe such moral weight to the humanity of the parties involved is not only in itself questionable, it also leads to the implausible conclusion that all sexual activity with non-human animals is morally permissible. To avoid this conclusion, the argument can be modified such that the apparent moral inconsistency between consentless sex with humans and non-human animals on the one hand, and consentless sex with robots on the other hand, is explained by the fact that the former are organic and the latter artificial. However, it seems to follow from the claim that the distinction between organic and artificial is morally relevant for the permissibility of consentless sex that the incorporation of artificial body parts into a human body, such as prostheses or 3D-printed organs, would diminish the moral standing of this human being. Moreover, if the appeal to the distinction between organic and artificial bodies is meant to indicate a morally significant difference between living and lifeless entities, life as such is ascribed intrinsic value, which has the counterintuitive implication that even the lives of bacteria must have such value.

Maybe the morally significant difference between sex without consent from a human being or non-human animal, and sex without consent from a robot, is that it harms the former but not the latter. An act harms a person if this person’s wellbeing decreases as a consequence of this act. Harm often results from the experience of pain, but this is not necessarily the case, as is clear from the fact that some individuals have a preference for violent sex, in which they engage voluntarily and wilfully. However, if an individual undergoes sexual activity that involves pain without having given consent, the experience of pain will harm this individual. Another way in which consentless sex is often harmful for humans is by violating the authority an individual has over her body and denying her autonomous will.

Since sexbots lack both qualitative experiences and moral autonomy, and they seem not to have any claim of authority over their bodies, it is true that they cannot be harmed by consentless sex. However, it does not follow from the fact that sex without consent is harmless that it is therefore permissible. For one thing, painless sexual activity with a non-human animal that lacks a claim of authority over its body, as well as an autonomous will, is morally wrong, even though the animal is not harmed by the activity. Similarly, instances of painless sexual activity that involve severely cognitively and physically impaired individuals, who lack authority over their body, as well as an autonomous will, do not harm these individuals, yet such activity is seriously wrong.

A more plausible argument to support the intuition that there is a moral inconsistency between consentless sex with humans and non-human animals on the one hand, and consentless sex with robots on the other hand, would be that in contrast to most non-human animals and human beings who have withheld consent, or are incapable of expressing consent, robots are devoid of the capacities necessary to consent. That is, their capacities are not just rudimentary, distorted, or underdeveloped, as is true of non-human animals, children and severely cognitively impaired individuals, but absent in actuality as well as in potentiality. However, since the moral permissibility of a sexual act depends on whether or not the parties involved are capable of withholding or giving consent, differences in cognitive makeup between different entities that fail to meet this threshold appear to be irrelevant.

Nevertheless, the fact that robots do not have the intellectual capacities necessary to consent does become morally significant if it is indicative of moral status. Moral status seems to derive from an entity’s sentience and sapience, which come in degrees and depend on an entity’s inherent features. Sentience is normally defined as the capacity to have qualitative experiences, such as pleasure and pain, and sapience is taken to be the capacity to enjoy a degree of psychological continuity, which originates from certain sophisticated cognitive phenomena such as self-awareness, rationality, and moral autonomy.[ii] Since sexbots of the near future will lack both qualitative experiences and the sophisticated cognitive capacities necessary for sapience, it follows that they will be devoid of moral status.

To say that sexbots will lack moral status is to say that they do not matter morally for their own sake, and that they are not a factor to be considered in moral deliberation. If sexbots do not matter morally for their own sake, they are not the kind of entities that we require consent from in order to do things to them. By contrast, individuals who are under high influence of alcohol or drugs, as well as severely cognitively impaired individuals, children, and non-human animals, do have moral status by virtue of their being sentient and, to different degrees, sapient. Because they have moral status, they are entities that we require consent from in order to engage in morally permissible sexual activity. Since they are in fact incapable of withholding or giving consent, we must refrain from having sex with them. In conclusion, then, the morally significant difference between consentless sex with humans and non-human animals on the one hand, and robots on the other hand, is that we require consent from humans and non-human animals but not from robots, because the former have moral status and the latter do not.

It must be kept in mind that this conclusion depends on the level of technological sophistication likely to be implemented into the design of near future sexbots. If robots acquire some degree of sentience or sapience in the more distant future, although not many scientists working on artificial intelligence and robotics think this a plausible scenario, they would in fact become entities that count morally in their own right. That is, if robots become sentient or sapient, they would constitute a new category of entities with moral status to be considered in moral deliberation, alongside the existing categories of human and non-human animals. In such a future situation, certain ways in which robots have been designed and used – including the design and use of sexbots – may on reflection appear to be seriously wrong.

[i] David Levy, Love + Sex With Robots, (London: Duckworth, 2009).

[ii] The view that moral status somehow depends on sentience and sapience is defended, amongst others, by Shelly Kagan, ‘What’s Wrong With Speciesism?’ in Journal of Applied Philosophy 33 (2016), 1-21, Jeff McMahan, The Ethics of Killing: Problems At the Margins of Life, (Oxford: OUP, 2002) and Peter Singer, Practical Ethics, (Cambridge: CUP, 1993). It should be noted, however, that these authors disagree strongly on issues such as hierarchy in moral status and the justifiability of speciesism.

Share on

8 Comment on this post

  1. Anders Sandberg and John Danaher have written here earlier about sex robots, and covered some of the more controversial aspects. This essay is more limited in scope , and concentrates on some more formal aspects of consent. Since the nastier (and embarrassing) aspects are also ethically relevant, we can’t leave them out of the discussion. It is also notable that this essay does not mention feminism, since almost all opposition to sex robots come from feminists.

    My main problem with this essay is, that it starts from the premise that consent is necessary for sex, but makes no attempt to explain or justify this. It simply says:

    … there is a general agreement in both common sense morality and ethical theory that sexual activity that takes place without consent from one of the parties involved is morally wrong, which is also reflected in the juridical characterization of rape as sex without consent.

    I don’t think there is such a consensus at all. If there was, we would all be signing consent forms before sex, but nobody does that. The legal characterisation of rape as ‘sex without consent’ is a shorthand, and that should have been made clear. English law, for instance, defines rape in this way:

    (1) A person (A) commits an offence if—
    (a) he intentionally penetrates the vagina, anus or mouth of another person (B) with his penis,
    (b) B does not consent to the penetration, and
    (c) A does not reasonably believe that B consents.

    Some states define rape in terms of the use of force or threat of force, rather than absence of consent. To my knowledge no state has ever required prior consent to be given to any sexual activity which is otherwise legal. (The law obviously will not require consent to acts which are illegal in themselves, because they are simply prohibited and that’s that).

    The law cannot in practice require prior consent, because it would have to specify the preconditions for, and means of, consent, and that is a minefield of ambiguity and potential bureaucratic intrusion into private life.

    At a guess, Romy Eskens thinks she does not have to justify the requirement for consent, because she can simply take it as a premise, and go on from there. However that is not the case. Since there is no apparent legal or ethical requirement that humans give prior consent to otherwise legal sexual acts, why should robots have to do that? And if they don’t have to do that, then why do we need to consider the ethical ‘technical issues’ – such as whether they possess autonomy or sentience?

    1. Mr. Treanor’s objection exhibits some conceptual slippage: he is focused on ‘consent forms’ and ‘prior consent’, but the law he quotes and the essay under discussion only mention ‘consent’. Sexual assault cases are notoriously difficult to prosecute, but since convictions do take place, clearly the law has other methods of evaluating consent than contracts signed in advance.

      1. There is no ambiguity or ‘slippage’. The law, in some countries, uses lack of consent to define rape, but it does not require consent. That may be difficult to understand for some people, but that is the law.

        Put plainly, many people think that before a man has sex with a women, she has to consent. That is simply not true. There is no such requirement in law, nowhere, so far as I know. We can be quite definite about this, because there are indeed institutions that have introduced a consent requirement, some universities in the United States for instance. We therefore know what a consent requirement looks like. A consent form is merely an example of a consent requirement, which has the advantage of being a record. (There are also consent apps, which claim to provide a record of consent).

        The point is that the law does not require any procedure. The law does not say that a man must have the consent of a woman, before he has sex with here. There is no legal requirement to obtain consent to sex, prior to having sex, or during sex. I cannot put it more plainly than that.

        Equally, if a man has sex with a woman who does not consent, then in several jurisdictions that constitutes rape. Yes, this is confusing, but that’s the way the law works in those countries. Perhaps only lawyers can understand the distinction – legal logic is often puzzling. Some laws are more precise, and speak of ‘against her will’ – a negative formulation which is not identical to lack of consent.

        One way around the confusion might be to use two separate terms – consent and permission. Permission implies a performative act, whereas both consent and lack of consent can be passive. Consent is a mental state, and can only be inferred. And that is precisely what the courts do in rape cases, they infer consent and lack of consent from the circumstances, as the circumstances appear from the evidence. That is ethically problematic, since it is ultimately guesswork. In jurisdictions where rape is defined by use of force or threat, things are clearer: in principle there can be evidence of the act of force, or pronouncing a threat.

  2. It’s the dumbest rehearsal I read… Here the only one who can not differentiate between a maniqui and a person is who writes this article and who somehow supports it …In the hypothetical case that a sexual robot is created …. it is a machine, no more, no less, no feelings, no pain, no soul, no autonomy, so simple … to pose a moral dilemma for “Having sex” with a piece of plastic is tantamount to posing a moral dilemma for having sex with a toaster … and if hypothetically that supposed distortion of reality existed, Hollywood actors have long been turned psychopaths or serial murderers for spending a year or more trying to play one.

  3. “For an agent to exercise moral autonomy is for her to act on rules she has imposed on herself, to which end sophisticated intellectual abilities, such as consciousness, reason-responsiveness and future-orientedness, are required. ”
    I don’t think this is true.
    For example, let’s suppose there is a distant world, a Boltzmann planet, which is like Boltzmann brains, but stable and full of people. Those people did not give rules to themselves – they just came into being as a result of a quantum freak event in a sufficiently large universe -, but they are psychologically just like us. It seems apparent to me that they are free to exercise moral autonomy.

    I don’t think that a programmed must lack moral autonomy. In practice, they might lack it, because even if they’re very sophisticated, their psychological differences from us might make them akin to, say, very intelligent aliens from another planet who lack a moral sense, and perhaps (though this point is debatable), they would not be moral agents. But moral agents or not, they would be free to act, even if they’re not moral agents, so they can potentially give consent, just as a sufficiently smart alien from another planet could freely consent regardless of whether they happen to be moral agents – they would still be free agents.

    I don’t how the agents came into existence can preclude them from making free choices, if their minds are sufficiently complex. Of course, if they have no minds, that’s an entirely different story. But in that case, it’s not wrong to have sex with them for fun, at least if the person having sex with them knows they have no minds. If the human doesn’t know whether they’re freely consenting or are zombies (i.e., without the capability for subjective experience), I would say it would still be permissible to have sex with them as long as the human properly reckons that either their sexual partner consents, or their sexual partner is a zombie.

  4. This isn’t an argument. It’s women losing their role as the gatekeepers of sexual activity to robots. The insanity in the above article is an attempt at whatever works at keeping women in power; not an honest argument.

    1. Although I don’t believe that consent is possible, see my comments above, that does not imply that women in general have an obligation to have sex with men. We should recognise that the general position is, that men want more than women would prefer to provide, so there is an inherent scarcity. That does not however mean that women have ‘power’ over men, in the sociological or political sense. The mere possibility of sex robots, even if they are not available yet, makes such issues explicit, and it should also be an incentive to think about why humans have sexual reproduction in the first place, and what we ought to do about it.

      Romy Eskens, however, did not go into any of those issues. I find that disappointing, but on the other hand she is not defending women’s power or gatekeeping, as the last comment suggests.

  5. Do humans require consent from their refrigerators before they open the door? Do humans require consent from their alarm clocks before they set them for a particular time? Do humans require consent from their fan heaters before they turn them on or off?

    Robots are not “non-human animals”, they are 100% lifeless machines. The sex robots now being designed are simply upmarket plastic dolls, partially animated. They have nothing remotely resembling living cognition.

    This article is laughable tripe, and a sad testimony to how much of modern “ethical philosophy” remains hopelessly devoid of meaningful ethics.

Comments are closed.