Should We Campaign Against Sex Robots?

Abstract

In September 2015 a well-publicised Campaign Against Sex Robots (CASR) was launched. Modelled on the longer-standing Campaign to Stop Killer Robots, the CASR opposes the development of sex robots on the grounds that the technology is being developed with a particular model of female-male relations (the prostitute-john model) in mind, and that this will prove harmful in various ways. In this chapter, we consider carefully the merits of campaigning against such a technology. We make three main arguments. First, we argue that the particular claims advanced by the CASR are unpersuasive, partly due to a lack of clarity about the campaign’s aims and partly due to substantive defects in the main ethical objections put forward by campaign’s founder(s). Second, broadening our inquiry beyond the arguments proferred by the campaign itself, we argue that it would be very difficult to endorse a general campaign against sex robots unless one embraced a highly conservative attitude towards the ethics of sex, which is likely to be unpalatable to those who are active in the campaign. In making this argument we draw upon lessons from the campaign against killer robots. Finally, we conclude by suggesting that although a generalised campaign against sex robots is unwarranted, there are legitimate concerns that one can raise about the development of sex robots.

Full text

Should we campaign against sex robots?
John Danaher1, Brian D. Earp2, and Anders Sandberg3
1. National University of Ireland - Galway
2. The Hastings Center Bioethics Research Institute (Visiting Scholar)
3. University of Oxford, Future of Humanity Institute
Abstract: In September 2015 a well-publicised Campaign Against Sex Robots
(CASR) was launched. Modelled on the longer-standing Campaign to Stop Killer
Robots, the CASR opposes the development of sex robots on the grounds that the
technology is being developed with a particular model of female-male relations (the
prostitute-john model) in mind, and that this will prove harmful in various ways. In
this chapter, we consider carefully the merits of campaigning against such a
technology. We make three main arguments. First, we argue that the particular claims
advanced by the CASR are unpersuasive, partly due to a lack of clarity about the
campaign’s aims and partly due to substantive defects in the main ethical objections
put forward by campaign’s founder(s). Second, broadening our inquiry beyond the
arguments proferred by the campaign itself, we argue that it would be very difficult to
endorse a general campaign against sex robots unless one embraced a highly
conservative attitude towards the ethics of sex, which is likely to be unpalatable to
those who are active in the campaign. In making this argument we draw upon lessons
from the campaign against killer robots. Finally, we conclude by suggesting that
although a generalised campaign against sex robots is unwarranted, there are
legitimate concerns that one can raise about the development of sex robots.
Authors’ personal copy. Forthcoming book chapter (penultimate draft). Please cite as:
Danaher, J., Earp, B. D., & Sandberg, A. (forthcoming). Should we campaign against sex
robots? In J. Danaher & N. McArthur (Eds.) Robot Sex: Social and Ethical Implications
[working title]. Cambridge, MA: MIT Press. Draft available online ahead of print at:
https://www.academia.edu/25063138/Should_we_campaign_against_sex_robots.
* Note: this is a draft paper; further revision may still be made prior to final publication.
Feedback at this stage is welcome.

2
1. Introduction
September 2015 saw the launch of the Campaign Against Sex Robots
(CASR).i Spearheaded by Kathleen Richardson of Leicester De Montfort University
and Erik Brilling of the University of Skövde, the CASR models itself on the longer-
standing Campaign to Stop Killer Robots, which was itself organized to preemptively
ban fully autonomous weapons. CASR opposes the development of sex robots (for a
sense of what we mean by this term, see the chapter in this volume by McArthur &
Danaher) on the grounds that they are ‘potentially harmful and will contribute to
inequalities in society’.ii The campaign’s founders believe that ‘an organized approach
against the development of sex robots is necessary’.iii
The campaign received considerable media attention after its launch.iv There
also appears to be some support for the view it espouses. David Levy, one of the
leading proponents of the development of sex robots (Levy 2007), was organizing the
2nd International Conference on Love and Sex With Robots prior to the launch of the
campaign. The conference was due to be held in Malaysia in November 2015, but was
cancelled shortly after the launch of the CASR as a result of opposition from
Malaysian government officials. Although there is unlikely to be a direct causal link
between the campaign and the cancellation, its cancellation does nevertheless
demonstrate an appetite and receptiveness to the position being advocated.
Given this apparent potential for the CASR to gain meaningful traction, it
seems important to assess its merits. That is what we attempt to do in this chapter. We
do so in two main parts. First, we try to examine the objections to the development of
sex robots that have actually been put forward by the CASR. We conclude that these
objections are ultimately unpersuasive, although we suggest that the degree of
(un)persuasiveness depends upon the actual aim of the campaign, which is currently
not well-specified. Second, we try to broaden the focus by asking whether there could
ever be a good reason to object to the development of something like this technology.v
To do this, we draw upon lessons from the Campaign to Stop Killer Robots. We
conclude that we are unlikely to come up with good reasons to preemptively reject the
development of sex robots, unless we adopt a highly conservative approach to the
ethics of sex, which many will find problematic for other reasons. This does not mean
that there are no legitimate concerns one can have about the development of sex

3
robots; it simply means that the concerns are not best addressed by adopting an
organized campaign against their development.
More generally, the analysis we present in this chapter has implications for
how societies should respond to controversial new technologies, based upon a range
of factors including risk (and type) of harm, prospect of benefit, and different ways of
pursuing regulation. Lessons learned here, therefore, should apply well beyond this
specific debate over sex robots and autonomous weapons.
2. What is the argument behind the Campaign Against Sex Robots?
We start by engaging the campaign on its own terms. To do this, we turn to
the position paper authored by Kathleen Richardson (2015).vi This paper sets out a
series of objections to the development of sex robots, which are grounded in an
analogy between sex robot-human relations and prostitute-johnvii relations. This way
of framing things suggests that the main intellectual grounding for the CASR (at least
in its current form) is an argument from analogy. Richardson makes this explicit at the
outset of her paper, commenting in particular on the model of human-robot relations
that is adopted in the work of David Levy (2007):
In his book, Sex, Love and Robots [sic]viii David Levy proposes a future of
human-robot relations based on the kinds of exchanges that take place in
the prostitution industry. Levy explicitly creates ‘parallels between paying
human prostitutes and purchasing sex robots’ [p.194] . I want to argue
that Levy’s proposal shows a number of problems, firstly his
understanding of what prostitution is and secondly, by drawing on
prostitution as the model for human-robot sexual relations, Levy shows
that the sellers of sex are seen by the buyers of sex as things and not
recognised as human subjects. This legitimates a dangerous mode of
existence where humans can move about in relations with other humans
but not recognise them as human subjects in their own right.
(Richardson 2015, 290)

4
Although Richardson does not spell out her objection to sex robots in formal terms,
we propose that it can be reconstructed in the following manner:
(1) Prostitute-john (or sex worker-client) relations are ethically problematic (for a
number of reasons, but particularly due to objectification of the sex worker).
(2) Sex robot-human relations are being modeled on sex worker-client relations
and so will share or exemplify similar properties to those relations.
(3) Therefore, sex robot-human relations will be ethically problematic (by
analogy).
(4) Therefore, we ought to campaign against sex robots.
This argument is not formally valid. Analogical arguments tend not to be (Walton
2013); they are informal defeasible arguments, dependent on the strength of the
similarity between the two cases being used. There is also something of a gap
between the first conclusion (3) and the second conclusion (4). Nevertheless, it is our
view that the reconstruction is fair to the position put forward in the paper by
Richardson because the prostitute-john analogy is front and centre in her analysis;
moreover, the gap between the first conclusion and the second conclusion is
something we wish to highlight in our critique.
Granting that this reconstruction fairly captures the argumentative centerpiece
of the CASR, we will now attempt to show that this argument suffers from three
major flaws, each of which serves to undermine the CASR in its current form.
2.1 – The first flaw: The objective is unclear
The first problem with the CASR’s argument is the gap between the two
conclusions. Why is it that ethically problematic properties in human-sex robot
relations support an organised campaign against the development of sex robots? The
inference, as stated, is unwarranted. Now, there may be a way to bridge this gap, but
the strength of the inference from (3) to (4) will depend on the objective of the
proposed campaign. Unfortunately, the CASR are not clear about their intended

5
objective. Until we have that clarity, it will be nearly impossible to tell whether or not
the campaign is something that is worthy of our support. Why is this?
Broadly speaking, when it comes to the development of sex robots (or any
potentially harmful technology), there are three main attitudes that we can adopt. We
can be entirely prohibitive in our outlook, favouring the preemptive suppression of the
technology. This is the attitude adopted by the Campaign to Stop Killer Robots in
response to the development of fully autonomous weapons systems.ix This prohibitive
attitude could extend to the criminalization of those who create, distribute and use the
technology (Danaher 2015), though criminalization may not be necessary.
Alternatively we can be entirely libertarian in our outlook, favouring complete
freedom in the development and use of the technology. Finally, in between these two
extremes, we can be regulative in our attitude, favouring some oversight and
intervention into the development and use of the technology, but falling short of a
complete ban. This regulatory attitude could take many forms, ranging from the
strong prohibition of certain types of sex robot while tolerating or even possibly
promoting others, to the weaker crafting of industry ethical standards, advertising
guidelines, safety protocols and the like.
In short, there is a spectrum of possible responses to the prospect of sex robot
development. The strength of the argument in favour of the CASR depends almost
entirely on where along this spectrum the objective for the campaign lies. Clearly, it
does not lie at the extreme libertarian end: there is something about the development
of sex robots that the CASR finds objectionable. But then where between total
prohibition and weak regulation does it lie? The answer makes a difference to the
burden of proof that the CASR’s argument needs to discharge. If they favour weak
forms of regulation – which, for example, might just ask manufacturers of these
robots to keep in mind various ethical issues that are raised by the technology, and
make some effort to address those issues – then the burden of proof will be relatively
low. Indeed, if there are plausible ethical concerns, then we could probably all agree
that these should be factored into the regulation of this technology. If, on the other
hand, they favour total prohibition, the burden of proof is much higher. In liberal
political regimes, there is (or should be) a general presumption against total
prohibition of new technologies, particularly when it comes to a technology designed

6
for personal use. The ethical issues identified by the campaign would need to be
extremely compelling to warrant preemptive suppression.
So where along the spectrum do the campaign’s objectives lie? Unfortunately,
this is unclear. The CASR’s website clearly states that an ‘organized approach against
the development of sex robots’ is required;x on their ‘About’ page they list several
unqualified objections to the development of sex robots; and in their position paper
and support materials they appeal to the preemptive model of the Campaign to Stop
Killer Robots. All of this suggests that their objective is close to the total prohibition
end of the spectrum. But there is also evidence of a more moderate view. The ‘About’
page, for example, states that the campaign is in favour of ‘ethical technologies that
reflect human principles of dignity, mutuality and freedom’ and that they want
‘computer scientists and roboticists to examine their own conscience when asked to
provide code, hardware or ideas to develop this field’. Richardson’s reliance on the
‘prostitute-john’ analogy in her position paper also suggests that if sex robots could
engender an alternative and more egalitarian relationship with humans then her
opposition might wane.xi Furthermore, in one media report, Richardson is recorded as
saying that we should simply ‘examine what it means’ to create such technologies,
not completely ban them; although in the very same piece she is also reported to have
called for a ban (Boran 2015). This suggests some equivocation on the part of the
campaign’s most visible founder as to its overarching objective.
This equivocation is problematic and may also explain some of the media
appeal of the CASR. As other chapters in this book make clear, ethical objections to
the development of sex robots are not uncommon. The fact that the particular set of
objections raised by the CASR has seen such widespread coverage is, at least prima
facie, puzzling. But now we may have an explanation: by ostensibly favouring the
extreme of prohibition, the campaign can generate much initial media interest, and
then when pushed on this extreme view it can retreat to a more modest and reasonable
position, a position that may not actually warrant an organized approach against the
development of sex robots (notwithstanding the stated aims of the campaign). If so,
this would constitute a so-called ‘motte and bailey’ tactic that is not uncommon in
debate (Shackel 2005). This is just a speculation. The more serious point is that until
the objective is clarified the CASR should not win widespread support.

7
2.2 – The second flaw: It paints a misleading view of sex work
The second problem with the CASR’s argument is its reliance on the sex
worker-client analogy. This analogy may not be essential to the campaign’s case – a
point to which we return below – but it features heavily in Richardson’s position
paper. Hence it is worth taking seriously. The claims that seem to bexii driving the
argument are that sex worker-client relationships have several bad-making properties
and that these bad-making properties will be shared and (thereby) reinforced and/or
normalised by sex robot-human relationships.
In advancing this position, Richardson adopts an almost entirely negative view
of sex work. Indeed, she objects to ‘sex work’ terminology in and of itself as well as
to related discourse, seeing it as an attempt to normalise sex work as a reasonable
choice in the labour market, when in fact, according to Richardson, it is highly
problematic. In support of this view, she cites various studies showing that
prostitution often occurs ‘in the absence of consent’, that violence and trafficking are
common in the industry, and that many of the workers are young girls or young
women (between the ages of 13-25). She does not engage at all with the position of
those who do not share her uniformly negative view of prostitution/sex work,
including some who have offered constructive articulations of alternative viewpoints
in the recent literature (e.g. Moen 2014). Instead, she seems simply to take for granted
that her position should be accepted as the only reasonable one that could be plausibly
advanced.
Her main objections to sex work are that it is built around an asymmetrical
power-relation between the client and the prostitute (one in which the client has the
balance of power); that the work itself is highly gendered (primarily women, and
some men, providing sex to primarily male clients);xiii and that the commercial
relationship involves the denial of subjectivity (and hence effective objectification) of
the sex worker. In other words, according to this view, the sex worker is treated as a
plaything that can be used and manipulated at the whim of the client, not as a human
being with feelings, agency and autonomy that ought to be respected. To back these
objections up she quotes from a study by Farley, Bindley and Golding (2009), which
included interviews with clients making claims such as these:

8
‘Prostitution is like masturbating without having to use your hand’,
‘It’s like renting a girlfriend or wife. You get to choose like a catalogue’,
‘I feel sorry for these girls but this is what I want’
(Farley et al 2009, 8)
Each of these views seems to reinforce the notion that the sex worker is being treated
as little more than an object and that her subjectivity is being denied. The client is
elevating his status and failing to empathise with the prostitute: he is substituting his
fantasies for her real feelings. Insofar as that is indeed what is going on in a client-sex
worker relationship, and if that is representative of such relationships, then we agree
that it would be a serious problem. Among numerous other reasons, the failure or
inability to empathise is often associated with higher rates of crime and violence
(Baron-Cohen 2011).
We thus seem to have two main lines of objection to sex work. The first is
concerned with violence and lack of consent within the industry; and the second is
concerned with objectification and the lack of empathy this entails. These are the bad-
making properties of sex work that support the first premise of Richardson’s
argument. Leaving aside the question of whether these bad-making properties carry
over to the case of sex robots, there are at least two problems with the objections
raised regarding the existence (and nature) of these bad-making properties that may
prevent the overall analogy from getting off the ground.
The first problem is that both objections are derived from a highly selective
view of sex work. Richardson paints a wholly negative picture of the industry. But
there are many prominent sex work researchers who challenge that picture. For
example, Sanders et al. (2009) provide an overview of the empirical literature on sex
work in which they develop a far more nuanced account. For instance, on the issue of
violence in the industry, they note that it is a ‘minority’ of clients who commit
violence, and that the vast majority of commercial transactions take place without
violence or other incident (Sanders et al 2009, 44), similar to the situation regarding
non-commercial sexual relationships. Likewise, on the question of objectification and
asymmetrical power relations, they note that much of this opposition is grounded in a
conservative view of sexual ethics that fails to consider the possibility of emotionally
rich and intimate sex occurring outside of the confines of a long-term monogamous

9
relationship (Sanders et al 2009, 83). Furthermore, they argue that the attitudes of
clients towards sex workers are often far more complex and multi-dimensional than
Richardson supposes:
…[G]eneral understandings of sex work and prostitution are based on false
dichotomies that distinguish commercial sexual relationships as dissonant
from non-commercial ones. Sanders (2008b) shows that there is mutual
respect and understanding between regular clients and sex workers, dispelling
the myth that all interactions between sex workers and clients are emotionless.
There is ample counter-evidence (such as Bernstein 2001, 2007) that indicates
that clients are ‘average’ men without any particular or peculiar
characteristics [who are] increasingly seeking ‘authenticity’, intimacy and
mutuality rather that trying to fulfil any mythology of violent, non-consensual
sex.
(Sanders et al. 2009, 84)
With respect to non-commercial sexual relationships, such as might exist between
‘ordinary’ couples (i.e., couples for whom the explicit exchange of money is not
directly associated with the sexual encounter, although other types of less overt
exchanges may sometimes occur), it must be acknowledged that these are neither
inherently nor exclusively models of ‘authenticity, intimacy and mutuality’ either,
notwithstanding their non-commercial nature. While this is arguably far from optimal,
what it suggests is that that the quality of sexual relationships in terms of respect,
power-asymmetries, etc., fall on a spectrum both within and outside of the
commercial context; hence, much more work would be needed to show that there is
something unique to an exchange-based sexual relationship that strictly entails the
undesirable qualities being discussed, and which does so in a way that is significantly
worse than comparable (i.e., casual) non-exchange-based sexual relationships.
We do not raise these points to paint a rosy and Pollyannaish view of sex work.
We are only trying to highlight the need for greater nuance than Richardson seems
willing to provide. It is simply not true that all forms of prostitution involve the
troubling features she identifies, while at the same time, many forms of non-
commercial sex do involve those features. Given that that is the case, the proposed

10
analogy with sex robots appears to be on much weaker initial footing than Richardson
assumes.
This brings us to the second criticism. Even if Richardson is right about both
(a) the bad-making properties of prostitution, and (b) the view that these properties
will “carry over” to the case in which the “prostitute” is a non-human robot (thus
potentially normalising the problematic attitudes that may be associated with such
bad-making properties; but see discussion below), there is no strong reason to think
that such an outcome would be sufficient to warrant a total prohibition on sex robots,
nor even, perhaps, an organized campaign to oppose them. For better or worse,
denials of subjectivity, asymmetries of power, and other objectionable qualities or
outcomes are rife throughout the capitalistic workplace. While this might form the
basis of a general critique of capitalism (such as a Marxist critique), it does not appear,
in most other cases, to motivate, much less successfully ground, a prohibition of the
activities or services that have such undesirable features. For example, many service
workers in our economies regularly have their subjectivity denied by their clients: one
may fail to care about the feelings of the massage therapist giving them a massage, for
example, or about those of the carpenter repairing their deck. But while we may have
very good reason to attempt to change the norms of behavior concerning such
uncompassionate treatment of service workers, it would seem a strange and rather
ham-fisted approach to attempt to ban the service industries altogether. Instead, a
change in regulation and social attitude may be all that is appropriately called for.
We must be clear that we are not suggesting that there are no morally-relevant
differences between the (typical) relationship between a client and her massage
therapist (or between a homeowner and her carpenter) and the (typical) client-
prostitute relationship. Instead, we are only pointing out that the presence of even
gravely undesirable features in a given type of exchange relationship (up to and
including denials of subjectivity and problematic asymmetries of power) is not
normally seen as being a sufficient justification for banning the relationship
altogether; rather, regulations that seek to mitigate the undesirable features that do
exist may worthily be pursued. On top of this, and perhaps even more importantly,
there is reason to think that many of the negative features of client-prostitute
relationships in particular could actually be directly caused (or worsened) by its
prohibition. This is because prohibition “cannot alleviate the background conditions

11
that contribute to people’s motivation to sell sexual services in the first place” (Earp
and Moen 2015, 54).
As one of us has argued elsewhere (drawing on the work of Luke Semrau
2015):
… there is an important distinction to be drawn between “being pressured to
sell X” and “being pressured, with the option to sell X.” If someone is being
directly pressured to sell [something] (or become a prostitute, etc.), then
prohibiting the selling of X could at least in principle be of some help. [But] if
someone is experiencing a more general pressure (such as extreme economic
insecurity), but has a number of ways—including, but not limited to, selling
X—to begin to relieve this pressure, then prohibiting the selling of X is
actually more likely, all else being equal, to make this person even worse off.
This is because it would remove (or drive underground, and therefore make
more dangerous) at least one otherwise viable option for “making ends meet.”
As a consequence, the person who was considering selling X, and who would
do so if it were not prohibited, must now turn to an even less desirable option
(as judged by them) to relieve the more general pressure.
(Earp and Moen 2016, 55)
Indeed, some of the problems of prohibition can arguably be seen in the case of
violence and trafficking. Insofar as prostitution is criminalized, sex workers cannot
avail themselves of the various protections that are afforded to most other kinds of
workers and they may also fail to report what happens to them for fear of adverse
legal consequences. This is why many sex worker activists — who are in no way
unrealistic about the negative features of the job — favour legalisation and regulation,
as opposed to outright prohibition (for further discussion, see Murphy 2015).
In sum, although there may certainly be harms associated with prostitution
that are different to and/or worse than the harms that are associated with other forms
of work, simply referring to those harms without showing how they are unique,
uniquely bad, or would be reduced rather than exacerbated by a ban, is not enough to
show that prohibition would be warranted.

12
2.3 – The third flaw: Application of the analogy is unpersuasive
Fortunately for Richardson, proving the badness of prostitution may not be
essential to her argument. All the CASR really needs to show is that sex robot-human
relationships will have bad-making properties which may be similar to (or intensified
versions of) the supposed bad-making properties of sex worker-client relations. If she
can show this, then the possible disanalogies between the two scenarios may not
undermine the case for organized opposition to the development of sex robots. Indeed,
it may be that these disanalogies support opposition in the case of sex robots.
One obvious disanalogy between the two cases is that sex workers are persons
and hence of high moral status; whereas sex robots are unlikely to be persons, at least
for the foreseeable future. This disanalogy can cut both ways. On the one hand, it
suggests that sex robots cannot be moral victims and hence we need not be concerned
about their treatment at the hands of their users; on the other hand, the fact that they
are not moral victims means that we don’t need to worry about the negative impact of
restrictive and prohibitive policies on their well-being (as we do with human sex
workers). This may cause us to favour an organized campaign against their
development, particularly if we think that (1) the harms arising from their use could
carry over to real human victims (through normalizing problematic attitudes or norms
of behavior), and (2) the putative benefits of their use are minimal.
This raises an important interpretive question about the intention behind
Richardson’s analogy with sex work. Is her claim that, just as the treatment of and
attitudes toward prostitutes is bad, so too will be the treatment of and attitudes toward
sex robots? Or is it that the development of sex robots will increase the demand for
human prostitution and/or encourage users of sex robots to treat real human (females)
more as objects? Richardson’s paper supports the latter interpretation. At the outset,
she states that her concern about sex robots is that they:
[legitimate] a dangerous mode of existence where humans can move about in
relations with other humans but not recognise them as human subjects in their
own right.
(Richardson 2015)

13
The key phrase here seems to be ‘in relations with other humans’, suggesting that the
worry is about how we end up treating one another, not how we treat the robots
themselves. This is supported in the conclusion where she states:
In this paper I have tried to show the explicit connections between prostitution
and the development and imagination of human-sex robot relations. I propose
that extending relations of prostitution into machines is neither ethical, nor is it
safe. If anything the development of sex robots will further reinforce relations of
power that do not recognise both parties as human subjects.
(Richardson 2015, S.4)
Again, the emphasis in this quote seems to be on how the development of sex robots
will affect inter-human relationships, not robot-human relations. Let us reflect this in
a modified version of premise (2):
• (2*) Sex robots will add to and reinforce the bad-making properties of
prostitution (i.e. they will encourage us to treat other humans with a lack of
empathy and thereby exacerbate existing, harmful gender/power inequalities).
Richardson supports this premise by focusing on the work of David Levy (2007), who
draws explicit parallels between the development of sex robots and prostitution.
Levy’s suggestion is that the relationship between a user and his/her sex robot can be
akin to the relationship between a client and a prostitute. Levy is quite explicit about
this and spends a good portion of his book looking at the motivations of those who
purchase sex and how those motivations might transfer onto sex robots. Interestingly,
Levy is far more nuanced in his discussion of this literature than Richardson is
inclined to be, highlighting some of the complexity in the motivations of clients and
also including a discussion of female clients of male sex workers (a perspective that is
entirely absent from Richardson’s analysis). In any event, the inference Richardson
draws from Levy’s work is that the development of sex robots is proceeding along the
lines that Levy imagines and hence we should be concerned about its potential to
reinforce the bad-making properties of prostitution.

14
Richardson identifies two major criticisms of her claim. The first, mentioned
above, holds that if robots are not persons then there is nothing wrong with treating
them as objects/things which we can use for our own pleasure. In other words, the
technology is a morally neutral domain in which we can act out our fantasies. The
second criticism points to the potentially cathartic effect of these technologies. If
people act out negative or violent sexual fantasies on a robot, they might be less
inclined to do so to a real human being. This suggests that sex robots may
consequently help to prevent the bad things that Richardson worries about. This view
is explored in more detail in other chapters in this book (McArthur, Strikwerda and
Danaher)
Richardson has responses to both of these criticisms. She argues that
technology is not a value-neutral domain. Our culture and our norms are reflected in
our technology. We should be worried about how cultural meaning gets incorporated
therein. Furthermore, she has serious doubts about the catharsis argument. She points
to the historical relationship between pornography and prostitution. Pornography has
now become widely available, but this has not led to a corresponding decline in
prostitution nor, in the case of child pornography, abuse of real children. On the
contrary, prostitution actually appears to have increased while pornography has
increased. The same appears to be true of the relationship between sex toys/dolls and
prostitution:
The arguments that sex robots will provide artificial sexual substitutes and
reduce the purchase of sex by buyers is not borne out by evidence. There are
numerous sexual artificial substitutes already available, RealDolls, vibrators,
blow-up dolls etc. If an artificial substitute reduced the need to buy sex, there
would be a reduction in prostitution but no such correlation is found.
(Richardson 2015)
Is this a robust defence of the bad-making properties of sex robot-human relations? Is
there enough here to warrant an organized campaign against their development?
We have some doubts. The evidence adduced to show that sex robots will
exacerbate harmful inter-human relationships is weak, and even if it is correct it does
not support a strongly restrictive approach to the development of sex robots. At best,
it supports a regulative approach (see analysis above). Furthermore, if we were to

15
adopt such a regulative approach we would need to be sensitive to both the merits and
demerits of this technology and the costs of any proposed regulatory strategy. This is
something that Richardson neglects because she focuses entirely on the negative. In
this vein, we offer four responses to her argument, some of which target her support
of premise (2*), others of which target the relationship between any putative bad-
making properties of sex robots and the need for an organised ‘campaign’ against
them.
First, Richardson’s primary support for premise (2*) - viz. that the prostitute-
john relationship is reflected in the model of sex robot development used by David
Levy — is weak. We have already noted that Levy is more sensitive to the complexity
of sex worker-client relations than Richardson is. But even if he weren’t, the
argument would not be very strong. True, Levy is a pioneer in this field and has a
degree of influence, but that doesn’t mean that all sex robot developers are obliged to
adopt his model. If we are worried about the relationship between the sex robot user
and the robot, we can try to introduce standards and regulations that reflect a more
positive set of sexual norms. For instance, the makers of Roxxxy (billed as the
world’s first sex robot) claim to include a personality setting called ‘Frigid Farah’
with their robot.xiv Farah will demonstrate some reluctance to the user’s sexual
advances. One could argue that this reflects a troubling view of sexual consent: that
resistance is not taken seriously (i.e. that ‘no’ doesn’t really mean ‘no’). Does this
mean that one should oppose the development of sex robots tout court? Not
necessarily. Instead, one could address the problem by regulating against this kind of
setting and insisting that every sex robot is required to give positive, affirmative
signals of consent. Although there are numerous problems that might be raised by the
prospect of an ever-willing (robot) sexual partner (see Danaher, this volume, for a
longer analysis), such an approach would at least plausibly reflect and reinforce a
more desirable attitude toward sexual consent. But we need not settle that example to
make the broader point: namely, that it is far from clear that there is anything inherent
in the nature of sex robots, qua sex robots, that would entail the promotion of negative
social attitudes toward women and sexual consent. Instead, we could demand and
enforce specific design regulations that would promote a more positive set of attitudes.
Maybe this is all Richardson really wants her campaign to achieve. But in that case,
she is not really campaigning against the sheer development of sex robots; she is

16
rather campaigning for a better version of them. If so, she could rename her campaign
‘the Campaign for Positive Sex Robots’, which changes the framing of the debate
considerably.
Second, we think it is difficult to make definitive claims about the likely link
between the use of a future technology like sex robots and actions toward real human
beings. Richardson’s claims about the correlation between pornography and an
increase in prostitution will be relatively unpersuasive to someone who does not think
that either pornography or sex work is unremittingly bad. What would be more
persuasive is if she could prove that there was some correlation (and ultimately some
causal link) between the increase in pornography/prostitution and the mistreatment of
sex workers. The evidence on this seems to be mixed. The legalisation of sex work is
sometimes associated with a decrease in mistreatment, and the related link between
pornography and sexual violence more generally is deeply contested. Some studies
show an increase;xv some show a decrease;xvi and some are mixed or uncertain.xvii
Danaher (2015) speculates that we may be landed in a similarly ambiguous position
when it comes to evidence concerning a link between sex robot usage and real-world
sexual aggression. Or, conversely, sex robots may be sufficiently different from
pornography as to increase the real-world effect. But in which direction is unclear: it
may embolden or satiate. It is simply too early and too difficult to tell. A
precautionary approach does not imply a ban as the first step taken, but rather a need
for more research and close observation (see Earp, Sandberg and Savulescu in press
for a defence of this view).
Third, and perhaps most importantly, when thinking about the appropriate
policy toward sex robots, or any new prospective technology, it is important that we
weigh the good against the bad. As two of us have argued previously (Earp et al. 2014,
7):
… when discussing the possible hazards associated with some predicted future
development—one has to remember that any new technology poses risks. This
is true whether it is an anti-love pill, a powerful military weapon, or
something more mundane. Hence the mere possibility that such a technology
might be used for ill can never by itself constitute sufficient reason to reject
it—however alarming such a possibility may be. Instead, the potential harms

17
that might accrue from the misuse of the technology must be weighed against
the potential benefits that might accrue from its responsible use.
In addition, as Bostrom and Roache (2011) have argued,xviii even the careful
anticipation of possible benefits and harms is not sufficient to give a full analysis of
the prudence of developing some new technology. Instead, these efforts must be
complemented with a meaningful attempt to identify “potential supporting policies
and practices that can alter the balance for the better” (144). As the philosopher C. A.
J. Coady has stated:
If indeed there is insufficient knowledge of outcomes and consequences, or no
social or institutional regulatory regime for prudent implementation of the
innovations and for continuing scrutiny of their effects, or no room for
overview of the commercial exploitation of the innovations, then … critics [of
new technologies] clearly have a point. [But] warnings can be heeded. [We
can] insist on safeguards and regulation, both scientific and ethical.
(Coady 2009, 165, emphasis added)
In the case of sex-robots, specifically, there are several potentially good-making
properties that would need to be factored into the discussion (McArthur, Di Nucci this
volume). For example, there is the simple hedonistic argument: sex robots provide
people with a way of achieving pleasurable states of consciousness. Whether this is a
good argument or not, or whether increasing hedonistic states should count for very
much when taken into consideration with other factors, is a complex question; but at
the very least, it needs to be engaged with in a serious way when evaluating the likely
value or disvalue for society of developing sex robots. Similarly, there is the
distributive argument: for whatever reason, there are people in the world today who
lack access to certain types of sexual experience, including people with certain kinds
of disabilities (see Di Nucci – this volume; see also Earp and Moen 2015) and sex
robots could make those experiences (or, at least, meaningful approximations of
them) available to such people. There is also the argument that sex robots could
ameliorate imbalances in sex drive between the partners in existing relationships
(McArthur – this volume), or could add a desired kind of diversity to the sex lives of
such couples, without involving human third parties (and the potential interpersonal

18
strife to which those third parties could give rise). It could also be the case that sex
robots would allow for particular forms of sexual self-expression and identity to
flourish (see generally Gupta 2012), including a pure sexual preference for machine-
sex (see Hauskeller 2014 for a longer, more critical discussion of this). If that turned
out to be the case, then, in the interests of basic sexual freedom and diversity, we
should permit it. This is by no means an exhaustive list of positive potential attributes.
It simply highlights the fact that there may be much good to the technology and this
must be weighed against any putative negative features when determining the
appropriate policy. Few new technologies are unalloyed goods; trade-offs must be
carefully considered.
Fourth, and finally, when thinking about the appropriate regulative policy one
also needs to think about the potential costs, both monetary and non-monetary, of that
policy. We might agree that there are certain bad-making properties arising from
human-sex robot relations, but it could be that any proposed regulatory intervention
would do more harm than good overall. Regulation of pornography, for instance, has
historically involved greater restrictions toward pornography depicting sexual
minorities (e.g., gay and lesbian porn).xix Regulatory intervention into sex robots may
end up doing the same, targeting robots used and designed by sexual minorities. It
may also be the case that policing the development and use of sex robots would
require significant resources and extensive intrusions into our private lives. Are we
willing to bear these costs? Less intrusive regulatory policies — e.g. ones that simply
encourage manufacturers to avoid problematic stereotypes or norms in the
construction of sex robots — would arguably be much more tolerable. Again, that
may be all that Richardson wants. But if that is the case, the campaign she is leading
needs to be much clearer about its aims.
2.4 – Interim conclusion
An interim summary is in order. The CASR’s primary argument against the
development of sex robots turns on an analogy between sex robot-human relations
and sex worker-client relations. The fear is that the former will reinforce the bad-
making properties of the latter. We find that this analogy does not support the
organized opposition to the development of sex robots, much less a full-fledged ban.

19
The analogy is premised on an overly negative view of sex work and the associated
need for its prohibition; its application to the sex robots scenario is weak and
speculative; and it also ignores the potential good-making properties of sex robots as
well as the costs of organized opposition. At most, we may have a basis for regulation
of sex robot development; but this seems to fall far short of the current rhetorical
demands of the campaign.
3. Should we ever campaign against sex robots?
The arguments of the CASR may be unpersuasive, but it is still worth asking
the broader question that their campaign might be seen to provoke: should we ever
campaign against a technology of this sort? This is a question of considerable social
importance. Artificial intelligence (AI) and robotics are on the rise. We can expect
more and more robots to enter into daily social use, including in ethically loaded
cases.xx Perhaps there are legitimate grounds on which we can oppose the
development of some of them. What is more, perhaps these grounds could apply to
the case of sex robots.
In this section, we address these issues by considering the arguments against
the development of fully autonomous weapons systems (AWSs), or ‘killer robots’.
Since the CASR is modeled after the Campaign to Stop Killer Robots, this seems an
obvious place to start. For context, AWSs are robotic systems that operate
independently of their human creators, administering lethal force to enemy targets.
There has been considerable social and academic opposition to the development of
such robots, going beyond the campaign just mentioned. For example there is the
Future of Life Institute’s (FLI’s) open letter on autonomous weapons systems, which
is strongly critical of the development of such technology.xxi Many people are
persuaded that AWSs simply should not be allowed to exist, and that a preemptive
ban on their development is in order. If arguments to that effect are indeed successful,
they might also apply to the case of sex robots in ways that have not yet been raised in
our discussion of the CASR.
In what follows, then, we survey some of the leading objections to the
development of AWSs, treating the issue largely on its own terms. After taking the
time to do so, we then return to the issue of sex robots, and conclude the chapter by

20
reflecting on some of the key lessons from this ‘killer robot’ analysis for the debate
over the CASR.
3.1 – Arguments opposing the development of AWSs
There are three main families of argument against the development of AWSs.
The first is pragmatically-oriented and is concerned with the consequences of
deploying AWSs in lieu of human soldiers. Proponents of this objection are
concerned with the practical limitations of the current and prospective technologies.
For instance, they worry about the targeting systems of AWSs and their ability to
adapt to dynamic battlefield conditions, in particular their ability to make fine-grained
and context-sensitive distinctions between who is and is not an enemy combatant
(Guarini and Bello 2012). The fear is that AWSs will be more indiscriminate in their
administration of lethal force than human operators would be and hence responsible
for worse outcomes. In making this argument, then, distinctions need to be drawn
between AWSs, which operate entirely independently of human controllers, and
teleoperated unmanned weapons systems, which are still under remote human control.
One can favour the latter on the ground that their use poses less risk to human soldiers
(Strawser 2010), while still opposing the former (e.g. Purves, Jenkins and Strawser
2015).
These pragmatically-oriented objections are important, and certainly fears
about targeting systems and battlefield adaptability will need to be addressed. But the
general utility of such objections is limited. As the underlying technology improves,
the consequentialist fears dissipate. Other objections to the development of AWSs try
avoid this technological contingency.
The second family of arguments is concerned with the need for moral agency
and responsibility in warfare. They can be grouped under the general heading of
‘responsibility gap’ arguments. The most-widely discussed example comes from
Robert Sparrow (2007). The gist of his argument is that the actions within a just war
must be carried out by, or ultimately be capable of being traced to, the decisions of a
responsible moral agent.xxii Humans are capable of being responsible moral agents.
AWSs are not. They lack the capacity for responsible agency and will always lack this

21
capacity. Hence, their use in a theatre of war opens up a responsibility gap: lethal
force is used but no one is ultimately responsible for its use.
Even if we grant the moral premise of this argument, we are still owed some
account of why it is that AWSs lack the capacity for responsible moral agency. Purves,
Jenkins and Strawser (2015) present two such accounts. The first focuses on the
codifiability of moral judgment. It maintains that AWSs will only be capable of
exercising moral judgment if moral judgment is capable of codification. AWSs will
have to be programmed and this programming will depend on well-defined rulesets
that identify the key moral requirements that the AWSs will need to follow. These
rulesets will then have to be reduced to a computer language.xxiii
The problem here is that several leading normative theories claim that moral
judgment is not ultimately codifiable in this manner. Seemingly robust moral
principles admit an indefinite number of exceptions that cannot be anticipated in
advance by human programmers; and some moral theorists claim that moral judgment
is always particularized (Dancy 2013). Accordingly, Purves, Jenkins and Strawser’s
second account of why AWSs lack appropriate moral agency focuses on other mental
constituents of responsibility. Following the views of leading just war theorists, it
argues that actors within a war are not simply required to act in conformity with
sound moral judgment; they must also act for the right moral reasons. In defending
this view, Purves et al. use the example of a racist soldier, who kills the right enemy
combatants but for the wrong reasons (Purves, Jenkins and Strawser 2015, 860). They
argue that the deployment of such a soldier would be wrong. The problem for AWSs
is that they will be incapable of acting for moral reasons. This claim builds upon
objections to strong AI (the belief that you could create an artificial being with
human-equivalent mental faculties). The leading accounts of what it means to act for
a reason require a sophisticated internal mental architecture that AWSs will lack for
the foreseeable future. Hence, they cannot act for moral reasons and so their
deployment in a theatre of war is contrary to the requirements of just war theory.
Both of these defences of the responsibility-gap argument can be challenged.
They rest on controversial metaphysical claims about the nature of mind and morality,
as well as the prospects for strong AI. Furthermore – as the authors themselves
acknowledge – important tradeoffs would arise if the robots achieved better battlefield
outcomes than human soldiers even while failing to meet the requirements for moral

22
responsibility. Others have used this possibility to develop defences of the use of
AWSs (e.g., Muller 2016; Simpson and Muller 2016). xxiv Nevertheless, if the
premises are sound, the arguments presented here provide strong moral reasons to
object to the use of AWSs that are not contingent on the current state of the
technology.
The final family of objections to AWSs focuses on the directness of the harm
they can potentially cause. This is perhaps the most significant consideration when it
comes to the merits of an organized campaign against their development. One thing
that all participants to the debate about AWSs can agree upon is the prima facie
wrongness of killing. In other words, everyone accepts that it is only permissible to
kill a human being in a limited range of circumstances. Killing someone when these
circumstances do not apply is a grave moral wrong. And even when the circumstances
do apply it is still bad for the (involuntary) victim. A key property of AWSs is that
their entire raison d’etre is to exert lethal force against enemy targets. In other words,
they are directly designed to do something that everyone agrees is a prima facie moral
wrong or a deeply regrettable outcome. The hope is that they will become
sophisticated enough to recognise when the circumstances that permit killing do in
fact arise. But there is always the risk that they will misjudge or miscalculate and do
something that is a serious moral wrong, or even enable moral wrongs on an
unprecedented scale (for example, enabling totalitarian states to harm or coerce
citizens with no possibility of soldiers’ consciences stopping them). This is one reason
why a preemptive campaign against the development of this particular kind of robot
seems so compelling: if the campaign succeeds, it can directly prevent a prima facie
wrong. We will return to this issue in our discussion of sex-robots near the end of this
chapter.
Now, some people object to this line of reasoning. They argue that the alleged
advantages of the preemptive ban are more illusory than real. AWSs are likely to be
developed by unscrupulous people, they claim, whether we try to prevent this
occurrence or not. A ban, they argue, would merely prevent research into making
them as ethical as is reasonably achievable. A related argument is that counterfactual
assessment of what would happen in the absence of AWSs is also important. Humans
are probably going to use lethal force against each other anyway just as they always
have. The advantage of AWSs, then, if they could be made sufficiently

23
technologically sophisticated, is that they could be more selective and less destructive
than human actors. But this line of thinking, too, can be challenged. One obvious
concern about the introduction of AWSs is that they can lower the threshold for
launching lethal attacks: if no human soldiers will be harmed in the process, a
commanding officer may be more likely to issue a command.
3.2 – Lessons for the Campaign Against Sex Robots
Now that we have some idea of the debate concerning AWSs, we should
reflect on the implications of that debate for the CASR. Some of the arguments we
have just surveyed seem relevant to the case of sex robots; others, less so. We proffer
no final judgment on the merits of the campaign against AWSs here (although we
note that one of us, Sandberg, has signed the FLI’s open letter against their
development; see above); instead, we simply wish to consider the lessons of that
campaign for our primary topic. To do this, we shall assume, arguendo, that at least
one of the preceding arguments reaches the threshold needed to warrant an organized
and systematic campaign against the development of killer robots. Do similar
considerations apply to the development of sex robots? We think not.
First, there are no equivalent practical concerns when it comes to the
development of sex robots. The practical fears surrounding AWSs comes from the
fact that they are intended to do harm and that inadequate programming or technical
sophistication could lead to the wrong kind (or degree) of harm. Sex robots, by
contrast, are not directly intended to cause harm; in fact, the opposite is true—they
are intended to cause pleasure. Certainly the robots could malfunction and these
malfunctions could give rise to health risks; or the robots could be used in dangerous,
perhaps unsanitary, ways. But these practical concerns seem comparatively minor, to
the point that they would not warrant preemptive organized opposition to the sheer
development of the robots in question. At most, they would warrant the creation of
appropriate industry standards as well as comprehensive user safety guidelines. Now,
it could be argued that the pleasure-inducing effects of sex robots could also be
addictive and hence a source of harm for those who become addicted to using them
(along with their families and friends). This is a more serious concern but the

24
potential harm is indirect and speculative. Hence we defer to our third response,
which we will come on to in just a moment.
Second, there is no equivalent need for moral agency and responsibility in the
design of sex robots. Again, the opposite would appear to be true. The need for moral
agency in the case of military robots stems from the morally fraught nature of their
actions and the need for moral accountability in warfare. But we don’t need sex robots
to be morally accountable. Indeed, turning sex robots into moral agents would
probably add to, rather than mitigate, ethical problems surrounding their use. For
instance, one of the putative benefits of sex robots (discussed by McArthur and Di
Nucci in this volume) is that they can facilitate positive sexual autonomy (i.e. desire
to have sex) without interfering with another person’s negative sexual autonomy (i.e.
desire to avoid sex). If the robot is a moral agent, then this putative benefit could
disappear. The robot would have to be given the same moral standing as any other
moral agent. Their consent to a sexual act would become a serious issue, and the
propriety of their use for sexual pleasure could no longer be taken for granted (cf. the
contributions from Petersen and Geras in this volume). This doesn’t seem like a
suitable ground on which to favour organized opposition to their development.
Third, there are no equivalently direct moral harms that could be prevented by
the campaign against sex robots. The harm from AWSs is direct and irreversible; the
alleged harms from sex robots are speculative, indirect and in principle reversible.
The argument from the CASR focuses on how the use of sex robots would impact on
real women (both those who work in the sex industry and those who do not). The fear
is that individual use of sex robots will distort the user’s downstream interactions with
real human beings and contribute to existing social problems arising from systematic
inequality and oppression of women. These are certainly legitimate ethical concerns,
and we agree that society should work hard to reduce systemic inequality, invidious
sexual objectification (cf. Marino 2008), and the harmful erosion of empathy. But
those problems stem from individual choices and social institutions, not from the sex
robots themselves. As one of us has previously put it:
…even if one bought the arguments that sex robots are likely to induce bad
social changes, these changes are occurring because of individual decisions
and beliefs, as well as sociocultural institutions. There are many other levers
that could be pulled to improve the situation of sex workers, women, or

25
people’s attitudes to each other. Some of these levers may be far more
powerful than a technology ban. Conversely, even a successful ban of sex
robots may fail to reach the desired goal because of other technologies or
intermediaries causing the undesired social changes. By acting against a
possible contributor rather than the bad thing itself, effort is wasted.
(Sandberg 2015)
In our view, this is the crucial difference with the campaign against killer robots. By
stopping killer robots you directly prevent a prima facie moral wrong. By stopping
sex robots, however, you do not.xxv
Direct harm arguments against sex robots are possible to make, but they are of
a very different kind to the ones used by the CASR. They would hold that there exist
inherent bad-making properties of human-robot sex that produce a direct—but
possibly victimless—moral harm. Arguments in this vein might fault human-robot sex
for being nonreproductive, for example, or for being a kind of inter-species sex
(assumed to be intrinsically wrong); for lacking essential relationship properties that
should exist during any sexual act (perhaps masturbation would be seen as wrong on
this view as well); or for being an immoral form of hedonism. Proponents of this view
could then argue that this direct and intrinsic (albeit, again, potentially victimless)
moral harm could be prevented from materializing if we were to pre-emptively ban
sex robots, just as preventing killer robots prevents a new moral harm.
Are such arguments convincing? To some people, they may be. However, they
appear to hinge on a peculiarly conservative and vice-related view of sex that is
unlikely to hold sway in many pluralistic societies. Thus, while certain individuals
may feel motivated to refrain from using sex robots (should they become available)
on the basis of these perceived direct moral harms, they would need to convince large
swaths of the population of their conservative view in order to justify a sweeping
change to policy. By contrast, AWSs are only legitimately going to be used by the
state, rather than by individual persons, for societal goals rather than personal goals.
That makes the rest of society proper moral stakeholders in decisions about their use
and development (since the state should represent and act in their interest), and
conversely they are to some extent responsible for what it does. That is not the case

26
with the personal use of sex robots, however, as the bedroom door insulates against
most would-be stakeholders.
Thus, while one can easily imagine social conservatives and vice-oriented
moralists supporting the CASR because of shared opposition to sex robots, they
would be doing so on the basis of a fundamentally different sort of moral claim – that
there is something inherently bad about human-robot sex – that would fit uneasily
with the stated ethics of the CASR’s leading proponents.
4. Conclusion
Robots are going to form an increasingly integral part of human social life.
Sex robots are likely to be among them. Though the proponents of the CASR seem
deeply concerned by this prospect, we have argued that there is nothing in the nature
of sex robots themselves that warrants preemptive opposition to their development.
The arguments of the campaign itself are vague and premised on a misleading
analogy between sex robots and human sex work. Furthermore, drawing upon the
example of the Campaign to Stop Killer Robots, we suggest that there are no bad-
making properties of sex robots that give rise to similarly serious levels of concern.
The bad-making properties of sex robots are speculative and indirect: preventing their
development may not prevent the problems from arising. Preventing the development
of killer robots is very different: if you stop the robots you stop the prima facie harm.
In conclusion, we should preemptively campaign against robots when we have
reason to think that a moral or practical harm caused by their use can best be avoided
or reduced as a result of those efforts. By contrast, to engage in such a campaign as a
way of fighting against—or preempting—indirect harms, whose ultimate source is not
the technology itself but rather individual choices or broader social institutions, is
likely to be a comparative waste of effort.

27
Notes
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
i For all documentation and supporting materials relating to the campaign, see
http://www.campaignagainstsexrobots.com/ (accessed 21/1/16)
ii Text taken from the CASR’s ‘About’ page:
http://campaignagainstsexrobots.org/about/ (accessed 21/1/16)
iii Text taken from the CASR’s ‘About’ page
http://campaignagainstsexrobots.org/about/ (accessed 21/1/16).
iv For a sample of the media coverage, see the ‘Press’ page on the CASR website.
v Here we follow the definitions of sex robot and having sex with a robot presented in
earlier chapters by McArthur and Danaher and Migotti.
vi The paper was published in the ACM SIGCAS newsletter, but a full text version can
be found at: http://campaignagainstsexrobots.org/the-asymmetrical-relationship-
parallels-between-prostitution-and-the-development-of-sex-robots/ (accessed 21/1/16)
vii Richardson prefers ‘prostitute-john’ to ‘sex worker-client’ because it is more
negatively value-laden. She disapproves of some attempts to create a positive
impression of sex work.
viii The actual title of Levy’s book is Love and Sex with Robots.
ix The campaign’s website states that a ‘A comprehensive, pre-emptive prohibition on
fully autonomous weapons is urgently needed’. See
http://www.stopkillerrobots.org/learn/ (accessed 21/1/2016)
x Text taken from the CASR’s ‘About’ page:
http://campaignagainstsexrobots.org/about/ (accessed 21/1/16)
xi Although, in saying this, we would need to know whether Richardson thinks that
this is possible. Some of her comments suggest that it wouldn’t be because robots
could never meet the criteria for personhood. See Richardson (2015).
xii The qualification ‘seems to make’ needs to be added because Richardson is never
that clear about her argument.
xiii Obviously, there are some sex workers providing services to women but they are
ignored in Richardson’s argument. This is not uncommon in the literature about the
ethics of sex work.
xiv The personality types are described on the ‘FAQ’ page on the TrueCompanion
website: http://www.truecompanion.com/shop/faq (accessed 4/5/2016)
xv See Malamuth and Chech (1981); Zillman and Bryant (1982 & 1989); and Allen et
al (1996).
xvi Ferguson and Hartley (2009); and Diamond (2009).
xvii Barak and Fisher (1997); and Barak, Fisher, Belfry and Lashambe (1999);
Malamuth, Addison and Koss, M. (2000); and Hald, Malamuth and Lange (2013).
xviii The rest of this paragraph is adapted from Earp, Sandberg and Savulescu (2014),
in which the quote following this paragraph also appears.
xix A famous example is the Canadian legal case of R v Butler in which the Canadian
Supreme Court proposed a new test for banning materials on the basis of obscenity. In
the aftermath of this decision, mainstream pornography continued to flourish while
pornography catering to minority sexualities was suppressed. For an extended
analysis see Crossman, Bell, Gotell and Ross (1997).
xx Examples could include autonomous cars which have to make life and death
decisions about whether to swerve to avoid an impact with a pedestrian or not; or the
darknet shopper, which is an automated shopping robot that randomly selects items
from the deep web for purchase. Given the nature of the deep web these can include
illegal drugs and firearms. For a description see: http://www.independent.co.uk/life-

28
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
style/gadgets-and-tech/news/random-darknet-shopper-exhibition-featuring-
automated-dark-web-purchases-opens-in-london-a6770316.html (accessed 17/2/16)
xxi This was signed by a number leading scientists, ethicists and robotics researchers.
See http://futureoflife.org/open-letter-autonomous-weapons/ (accessed 17/2/16)
xxii The argument appeals directly to the requirements of just war theory. There is
some dispute as to jus in bello (i.e., the justice of actions in a war) and jus ad bellum
(i.e., the justice of the decision to go to war) requirements and whether they offer a
way round the responsibility gap argument. Purves, Jenkis and Strawser (2015) argue
that this distinction should make no difference and, for the purposes of this discussion,
we concur.
xxiii There may be ways to create an AI that avoid codifiability. Machine learning
techniques show some promise in this respect. They can be given examples of moral
and immoral behavior and then extrapolate to new cases. They can sometimes learn
subtle exceptions, but the problem with machine learning is that it tends to be opaque.
We will not know whether an AI has figured out an important exception until we
place it in that situation: we cannot just assume that it has from proper behavior in the
training cases, and seeing proper behavior in other test cases merely gives us
inductive reasons to think it will behave appropriately in real-world situations. If the
exception is of a different type than the other test cases, then we have no real reason
to trust the system.
xxiv It is worth noting that in these papers, Muller and Simpson develop a position that
is similar to the one advocated in the present chapter, namely: there are bad-making
properties of AWSs that could warrant regulation and intervention, but not a
preemptive campaign against their creation.
xxivi A possible argument might be that if a person knew that nobody would be harmed
by having sex of some (inherently) problematic kind, they might do it more often. Sex
that is problematic because of harm to partners would now be unproblematic, but if
there is an inherent harm (as per the conservative argument) or indirect bad effects
(closer to Richardson’s view), then there is now more harm overall due to the actions
of these people. However, this would be overshadowed by the different actions of
other people: moral people who abstain from problematic sex because it is morally
problematic are likely a far smaller group in most cases than people who abstain from
it simply because it is hard to find partners.

29
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Bibliography
Allen, M., Emmers, T., Gebhardt, L. and Giery, M (1996). Exposure to pornography
and acceptance of rape myths. Journal of Communication 45: 5–26.
Barak, A. and Fisher, W.A. (1997). Effects of interactive computer erotica on men’s
attitudes and behavior toward women: an experimental study. Computers in Human
Behaviour 13: 353–369
Barak, A., Fisher, W.A., Belfry, S. and Lashambe, D. (1999) Sex, guys and
cyberspace: effects of internet pornography and individual differences on men’s
attitude toward women. Journal of Psychology and Human Sexuality 11: 63–91
Baron-Cohen, S. (2011). Zero degrees of empathy: a new theory of human cruelty and
kindness. London: Penguin.
Bernstein, E. (2001). The meaning of the purchase: desire, demand and the commerce
of sex. Ethnography 2(3): 389-420.
Bernstein, E. (2007). Sex work for the middle classes. Sexualities, 10(4): 473 – 488.
Boran, M. (2015, November 4). Sex robots should be banned, expert tells web summit.
The Irish Times. Retrieved from: http://www.irishtimes.com/business/technology/sex-
robots-should-be-banned-expert-tells-web-summit-1.2416948 (accessed 21/1/2016)
Bostrom, N. and Roach, R. (2011). Smart policy: cognitive enhancement and the
public interest. In , Savulecu, J., ter Meulen, R., and Kahane, G. (eds) Enhancing
human capacities, 138–152. Oxford, UK: John Wiley & Sons.
Coady, C. A. J. (2009). Playing god. In Savulescu, J., and Bostrom, N. (eds) Human
enhancement, 155–180. Oxford, UK: Oxford University Press.
Crossman, B., Bell, S., Gotell, L. and Ross, B. (1997) Bad attitudes on trial:
pornography, feminism and the Butler decision. Toronto: University of Toronto Press.
Danaher, J. (2015). Robotic rape and robotic child sexual abuse: should they be
criminalised? Criminal Law and Philosophy, in press. doi: 10.1007/s11572-014-9362-
x
Dancy, J. (2013). Moral particularism. In Zalta, E. (ed) The Stanford Encyclopedia of
Philosophy. Retrieved from: http://plato.stanford.edu/entries/moral-particularism/
(accessed 18/1/15).
Diamond, M. (2009). Pornography, public acceptance and sex related crime: a review.
International Journal of Law and Psychiatry 32: 304–314.
Earp, B. D., Sandberg, A. and Savulescu, J. (2014). Brave new love: the threat
of high-tech “conversion” therapy and the bio-oppression of sexual minorities. AJOB:
Neuroscience 5(1): 4-12.

30
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Earp, B. D., Sandberg, A. and Savulescu, J. (in press). The medicalization of love:
response to critics. Cambridge Quarterly of Healthcare Ethics: Clinical Neuroethics,
in press.
Earp, B. D. and Moen, O. M. (2015). Paying for sex—only for people with
disabilities? Journal of Medical Ethics 42(10): 54-56. doi:10.1136/medethics-2015-
103064
Farley, M., Bindel, J. and Goulding, J. (2009). Men who buy sex: who they buy and
what they know. London: Eaves
Ferguson, C. J. and Hartley, R. D. (2009). The pleasure is momentary ... the expense
damnable? The influence of pornography on rape and sexual assault. Aggression and
Violent Behavior 14: 323–329.
Guarini M. and Bello, P. (2012). Robotic warfare: some challenges in moving from
noncivilian to civilian theaters. In Lin, P., Abney, K. and Bekey, G. A. (eds) Robot
ethics: the ethical and social implications of robotics. Cambridge, MA: MIT Press.
Gupta, K. (2012). Protecting sexual diversity: rethinking the use of
neurotechnological interventions to alter sexuality. AJOB Neuroscience, 3(3), 24-28.
Hald, G. M., Malamuth, N. M. and Lange, T. (2013). Pornography and sexist attitudes
among heterosexuals. Journal of Communication 63: 638–660
Levy, D. (2007). Love and sex with robots: the evolution of human-robot
relationships. New York: HarperCollins.
Malamuth, N. M. and Chech, J. V. P (1981). The effects of mass media exposure on
acceptance of violence against women: a field experiment. Journal of Research in
Personality 15: 436–446
Malamuth, N. M., Addison, T. and Koss, M. (2000). Pornography and sexual
aggression: are there reliable effects and can we understand them? Annual Review of
Sex Research 11: 26–91.
Marino, P. (2008). The ethics of sexual objectification: autonomy and consent.
Inquiry, 51(4): 345-364.
Moen, O. M. (2014). Is prostitution harmful? Journal of Medical Ethics 40(2): 73-81.
Muller, V. (2016). Autonomous killer robots are probably good news. In Di Nucci
and Santoni De Sio (eds) Drones and responsibility: legal, philosophical and socio-
technical perspectives on the use of remotely controlled weapons. London: Ashgate.
Murphy, C. (2015, August 14). Sex workers’ rights are human rights. Amnesty
International. Retrieved from https://www.amnesty.org/en/latest/news/2015/08/sex-
workers-rights-are-human-rights/.

31
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Purves, D., Jenkins, R. and Strawser, B. (2015). Autonomous machines, moral
judgment, and acting for the right reasons. Ethical Theory and Moral Practice 18(4):
851-872.
Richardson, K. (2015). The ‘asymmetrical’ relationship: parallels between
prostitution and the development of sex robots. SIGCAS Computers & Society 45(3):
290-293. Retrieved from: http://campaignagainstsexrobots.org/the-asymmetrical-
relationship-parallels-between-prostitution-and-the-development-of-sex-robots/
(accessed 21/1/2016)
Sandberg, A. (2015, September 16). Sex and death among the robots: when should we
campaign to ban robots? Practical Ethics Blog. Retrieved from:
http://blog.practicalethics.ox.ac.uk/2015/09/sex-and-death-among-the-robots-when-
should-we-campaign-to-ban-robots/ (accessed 21/1/2016)
Sanders, T., O’Neill, M. and Pitcher, J. (2009). Prostitution: sex work, policy and
politics. London: Sage.
Sanders, T. (2008). Male sexual scripts: intimacy, sexuality and pleasure in the
purchase of commercial sex. Sociology 42(1): 400-417.
Semrau, L. (2015). The best argument against kidney sales fails. Journal of Medical
Ethics, 41(6), 443-446.
Simpson, T. and Muller, V. (2016). Just war and robots’ killings. The Philosophical
Quarterly 66(263): 302-322.
Shackel, N. (2005). The vacuity of postmodernist methodology. Metaphilosophy
36(3): 295-320.
Sparrow, R. (2007). Killer robots. Journal of Applied Philosophy 24(1): 62-77.
Strawser, B. (2010). Moral predators: the duty to employ unmanned aerial vehicles.
Journal of Military Ethics 9(4): 342-368
Walton, D. (2013). Analogical arguments. In LaFollette, H. (ed) The International
Encyclopedia of Ethics. Hoboken: Wiley-Blackwell.
Zillman, D. and Bryant, J. (1982). Pornography, sexual callousness and the
trivialization of rape. Journal of Communication 32: 10–21
Zillman, D. and Bryant, J. (1989). Pornography and men’s sexual callousness toward
women. In Zillmann, D. and Bryant, J.(eds) Pornography: research advances and
policy considerations. Hillsdale, NJ: Erlbaum.