The opinion of…

Dr Kate Devlin is a Senior Lecturer in Computing at Goldsmiths, University of London, and integrates the committee of the Third International Congress on Love and Sex with Robots, which will be held on 19-20 December, 2017 in London, UK.

Twitter: @drkatedevlin

May 4th 2017

Recent years have seen a rise in interest of the “sex robot” – a machine, often humanoid (or, more specifically, gynoid) to take the place of a human partner in sexual activity. This has generated considerable controversy – and rightly so. With the introduction of any new technology we should examine the longterm impact. My stance is that there are a number of issues that require research and exploration but that overall, and – importantly – with a directed and ethical approach, the benefits of such robots would outnumber the risks.

Their imagined history is long. Sex toys have been in use for hundreds, if not thousands, of years. The use of sex dolls has been recorded as far back as the seventeenth century. Sex robots (which are, to my mind, a more embodied form of the sex toy and a more sophisticated form of the sex doll) have long been conceptualized – from the myths of the ancient Greeks to the sci-fi of the current era. The trope of the sex robot is that of a posable, silicone woman: hyperfeminine, pornified and biddable. I argue that this is an unnecessary and potentially a damaging portrayal, and that we should move towards a fairer, more diverse imagining of the sexual companion robot.

The word “robot” has its roots in the Eastern European term for servitude, and was specifically used by the Hungarian playwright Karel Čapek in 1922 in his work R.U.R. (Rossum’s Universal Robots) to describe humanlike, manufactured people. He attributes his use of the phrase to his brother, the writer and artist Josef Čapek. In Karel’s play, the robots are incapable of original thinking but, like all good dystopian fictions, some of them gain self- awareness and rise up against their human creators.

The reality is that we already have robots in our lives today. They are machines designed to automate a task. We see them at use in the home (robot vacuum cleaners), in hospitals (robotic surgery), in the military (for bomb disposal, or for automated weapons systems), or as companions (such as therapy robot pets). None of these examples take a human form, though some have zoomorphic or anthropomorphic characteristics. In fact, we are a very long way from the convincingly human robot. The Uncanny Valley, a term coined in 1970 by Masahiro Mori, is a hypothesis that we humans empathise with machines that have human attributes, up until the point where they approach indistinguishability from reality. At that point, the “almost-human” robot becomes unsettling, uncanny and fills us with revulsion. There are a number of theories as to why this might be, including the idea that “human-looking but not alive” is redolent of death, or that the conflicting perceptual cues cause cognitive discomfort. We are still stranded in the uncanny valley. Current human-like robots could never be mistaken for a real person.

Robots are programmable: we, the developers and users, instruct them in the task they are to complete. In some instances, artificial intelligence is used to enable machines to perform tasks with a degree of independence in order to achieve a goal. Artificial intelligence is the concept of machines being able to carry out tasks in an intelligent manner. The key objective here is reaching the goal – not how that goal is reached, although varying branches of the discipline deal with the ways this might be achieved, e.g. by mimicking human cognition, or by taking a purely computational, mathematical approach. At present, none of the artificial intelligence we see in use today involves a machine actually being sentient or conscious. Rather, they can use data to understand patterns of outcomes and thus learn from previous situations. Currently, artificial intelligence is fairly good at handling narrow, domain-specific tasks (such as digital assistants like Amazon Echo or Apple’s Siri). General artificial intelligence, where a computer would need reasoning and commonsense knowledge to solve a real-world, unexpected problem, is a much, much more difficult aim.

Self-awareness is at the heart of AI research. Can we create a machine that can think? Could a computer have a consciousness? The jury is out. Some researchers, vastly aware of the human limitations in understanding our own cognition and perception, insist that it will never happen. Others feel that it’s possible – but it may not resemble our own recognisable form of consciousness. Robots can be a very specific type of artificial intelligence – one that exists in physical form. We do not have robots that can behave fully autonomously. We do not – and may never have – robots that are conscious.

In terms of sex robots, the state-of-the-art is the RealDoll and similar products: a life-size mannequin with limited interactivity. As of 2017, RealDoll creators Abyss Creations, are introducing artificial intelligence into their dolls. These dolls are hypersexualised representations of women – objectifiable and, on that basis, objectionable. It has been claimed that these dolls are a threat to women as the men who buy them can inflict violence on them without repercussion, and that this would normalise violence against women. However, many owners of the dolls (self- defined as “iDollators”) are actually strongly invested in their care. They report that they choose a relationship with a doll for therapeutic reasons, including lack of intimacy with their human partner due to physiological reasons such as a spouse’s chronic illness. Others prefer the safety of a doll following psychological problems resulting from the collapse of previous relationships.

The possibility of child sex robots is often given as an argument against the development of sex robots, but it must be noted that child sex dolls are already illegal – as is any computer generated imagery of child abuse (a court case in Canada is currently deciding whether or not a child-like sex doll constitutes child pornography. While the idea of a child sex robot is abhorrent, there is a need for investigation as to its therapeutic potential, particularly as an outlet or proxy for pedophiles in order to restrict offences to the virtual realm. As with the debate around the impact of pornography (not dissimilar to the research on violence linked to video games), the jury is out. Each side can show evidence that either exposure is damaging or, conversely, that exposure is beneficial. The “gateway” vs. “reduction” theories remain inconclusive and they bear investigation. Interestingly, the University of Montreal has trialed the use virtual reality (VR) in monitoring sex offenders – and there have been successful experiments where VR has been used to reduce psychological conditions such as social anxiety and phobias. If VR can make a difference, could a physical but non-real substitute also have influence?

Companion robots – non-sexual – are already in use today and the advantages and disadvantages of these are the focus of academic research. Extending this research into sexual companion robots is clearly a further step. The technology exists and the sex robot is becoming a reality – and a commercial viability. Rather than railing against them, we need to accept that this is happening and we should use the opportunity to shape their form and future. Better that we work now to influence the direction from an ethical and informed standpoint than to impose moralistic judgment that stigmatizes sex outside of a monoheteronormative relationship.

The gendered, stereotyped and clichéd options we have today don’t have to dictate that this what a sex robot must look like. In fact, we have wonderful materials, technologies and opportunities that move us into new and much more interesting territory. Advances in human-computer interaction mean we can communicate with technology via touch, speech, gesture – and even our brain waves. We can stream data from our bodies to give us instantaneous readings of our skin responses, heart rate, muscle movement and facial expressions. We have at our disposal a wonderful and exciting range of smart fabrics, conductive paint, soft robotics, and sensors. We already create robots that are not intended to be realistically human or gendered. Why not take this same approach with sex robots? Fundamentally, sexual activity (be it alone or with others) is associated with positive wellbeing measures. The use of sex technology – sex robots, sex toys, apps and digital content – is emerging as a innovative approach to enabling people to have an independent and fulfilling sex life where physiological, psychological and discriminatory barriers currently exist. To campaign against development is shortsighted. Instead, we should be advocating research into an area long regarded as socially taboo.