One researcher on why the oldest profession could survive (and perhaps thrive in) technological unemployment.
Image: Flickr/Caroline Davis2010. Below, inset: John Danaher.
As every year goes by, robots are becoming more useful to humans. They’re doing jobs that are too dangerous for us, taking care of our elderly, and even helping give us part of our bodies back. But robots are also developing to fulfill our more primal urges in the form of fully functioning sexbots—something that's already edging it way into reality.
Ethically, the concept of robosex is a somewhat contentious subject. On the one side, some robotics researchers say sexbots could be a boon to certain people, like those who have a disability that gets in the way of achieving physical relationships. They further argue that sexbots could do a lot of good by disrupting the human prostitution industry. But on the other side, some purveyors of social morality (and those who generally think people spend too much time with their smartphones already) fear that sex with machines will lead to the breakdown of human-to-human bonding on any level.
But one scholar from the Institute for Ethics and Emerging Technologies is throwing a spanner into both those arguments. John Danaher holds a PhD from University College Cork and is a lecturer in law at Keele University. He’s recently published a controversial paper in which he argues that sexual robots could actually promote human-to-human relationships. And human prostitution? That could be the hottest gig in town after the robots have taken all our other jobs. I talked to John Danaher at length about everything from the legal issues around robosex to robots displacing human jobs—and where regular prostitution will fit in. Here are the highlights of our discussion.
Motherboard: Many working in the fields of AI and robotics say a future of sexbots is inevitable—and even preferable to human prostitution. What are some of the advantages of sexbots over human prostitutes?
John Danaher: There are several potential advantages of realistic sexbots over human beings. In my paper, I break them down into four categories. First, sexbots might have legal advantages over humans. In many countries prostitution and its related activities (e.g. advertising, creation of brothels) are illegal. Consequently, sex workers and their clients risk legal liability every time they engage in a transaction. At present, sexbots do not face the same kinds of legal restrictions. That said, this advantage may be temporary. Many Western countries have liberalised their prostitution laws, and as the sexbot industry grows we can expect greater legal regulation and control of it.
Another advantage of sexbots over humans might be in relation to moral and social attitudes. Many people worry about the ethical aspects of human prostitution. Those worries may not arise when it comes to sexbots (though this depends on whether robots can have moral status). A third advantage could be in relation to the risks of contracting a disease. Sexual contact between human beings can, of course, carry the risk of contracting an STD. Provided there is proper sanitation, this risk can be minimised in the case of sexbots. A final advantage of sexbots could be their flexibility and ease of production. Sexbots could be programmed to cater to the demands of their users, they could be produced en masse and, like other industrial robots, they could work tirelessly for no pay.
"Another possible advantage of robots over humans is that they could do things that no human being could do."
I should enter two caveats about all these alleged advantages. The first is that it's not clear that these advantages would be sufficient to render human prostitutes obsolete: humans may have other advantages over robots that allow them to be competitive. The second is that these advantages relate to "realistic" robots, by which I mean "realistically human-like." Another possible advantage of robots over humans is that they could do things that no human being could do, make possible new non-human types of sexual experience. That's something that many futurists and transhumanists are keen on.
No one wants to see robots take away jobs from humans, until you talk about the profession of prostitution—then people’s tone changes. If sexbots could render human prostitution obsolete, would that be a good thing?
That depends on your attitude to prostitution. If you think prostitution is a terrible evil, that it does nothing but exploit and objectify prostitutes, then perhaps it would be a good thing. But if you think there is nothing inherently wrong with commercialised sex, maybe not. In that case it would be like any other form of technological unemployment. Is it a bad thing that factory workers lose their jobs to robots? That depends. Does it free them to do other more valuable things? Are their basic needs going to be provided for by some system of social welfare while they retrain? If so, then technological unemployment might be a good thing. But one of the worrying aspects of the current predictions about technological unemployment—as set out in, say, The Second Machine Age by Brynjolfsson and McAfee—is that these other more valuable opportunities may also be taken over by robots.
Contrary to others, you actually hypothesize that human prostitution, unlike other industries where robots will take over our jobs, is likely to remain resilient to technological unemployment even if realistic sexbots become available. You term this the “resiliency hypothesis.” Can you explain what that is exactly?
The resiliency hypothesis is the claim that prostitution may be one of the few areas of human labour that is resilient to technological unemployment. By resilient, I don't mean that sex robots won't be used; just that they won't replace human prostitutes. An analogy with sport might be helpful. I can imagine a day when highly realistic human-like robots could battle it out in televised fighting contests (indeed, televised robot fighting contests already take place). I think people could be interested in watching those contests. But I don't think that robotic fights will replace or overtake the interest in human fighting contests. People will still be more interested in human boxing and MMA and the like because they are more interested in the competition between human abilities. I think something similar could be true of the relationship between robot sex and human prostitution.
"Given a choice, most humans will prefer to have sex with another human than with a robot."
I defend the resiliency hypothesis on two main grounds in the paper. The first holds that there will be a consistent preference for human-to-human sex over human-to-robot sex. Given a choice, most humans will prefer to have sex with another human than with a robot. The second ground is that technological unemployment in other industries may actually increase the number of people working in the sex industry, thereby allowing them to out-compete the robots.
I don't fully endorse the resiliency hypothesis in my paper, but I think it is plausible, possibly more plausible than the claim that sex robots will displace human prostitutes.
Humanity’s ontological preferences for human-to-human sex is related to a phenomenon known as the uncanny valley effect, which you say will keep humans preferring humans no matter how lifelike sexbots get, correct?
The uncanny valley effect was originally proposed by the Japanese roboticist Masahiro Mori. His suggestion was that as robots became more human-like, a point would be reached where actual humans found them to be pretty creepy. This is because they would be too close to an actual human for comfort (too uncannily human-like). A classic example of the uncanny valley effect is said to be the reaction of viewers to the film The Polar Express. People found the CGI characters in that film to be too realistic for comfort. There has been a reasonable amount of research into this effect in recent years, with some of it even suggesting that other animals experience it too.
In terms of my paper, the relevance of this is obvious enough. I claim that the uncanny valley effect could put people off robot sex partners. This supports the human preference thesis and the resiliency hypothesis as a whole. Of course, as originally presented by Mori, the uncanny valley effect was just a "valley." He believed that once robots became pretty much perfect facsimiles of real human beings, people would no longer be creeped out. I review some recent studies suggesting that this may not be entirely true, and that the off-putting effect could be more difficult to overcome.
What’s even more interesting about humans of the future having a choice between human prostitutes and sexbots is the fact that you believe that with the rise of robots taking over our other jobs it will actually spur a new influx of humans into prostitution. Can you explain how you came to this conclusion?
This is my argument for the increased supply thesis. It's pretty simple really. If we accept that technological unemployment will increase across a range of industries—as many economists are now predicting—then all the people who are unemployed will be looking for other sources of income. Provided that there is no dramatic change to social welfare, this means they will have to look for other work. But where will they go if most jobs have been taken over by robots? Well, they'll go to those industries where there is a preference for human labour over robotic labour, where there are low barriers to entry (i.e. no significant qualification or retraining required), and which are comparatively well-paid. I argue that prostitution could satisfy all three of those conditions, thereby increasing the supply of human prostitutes.