Robots for sexual relations might seem just round the corner, but as Rose Eveleth argues, the reality is complicated.
In two recent movies, Her and Ex Machina, filmmakers explored an intriguing concept: whether humans will fall in love with and want to have sex with AI or robots. Of course, this isn’t the first time people have lusted after artificial human entities – you can find that story all the way back to the Greek myth of Pygmalion. But as AI and robotics get more and more advanced, some have claimed that robots designed for love and sex are finally close to reality. In his book Love and Sex with Robots, for instance, David Levy estimated that marrying robots will be normalised by 2050.
Yet while it might sometimes seem like we’re on the cusp of a new sexual revolution, the truth is more complicated. Building and designing robots for sex is going to be harder than most people realise, making them convincing without being creepy will be a huge hurdle, and overcoming the barriers the sex industry faces for funding will be enormous. The idea that a company might come along any day now and make an affordable and convincing sex robot ignores the reality of both research and regulation.
Before exploring why, however, let’s be more specific about what makes something a sex robot. Technically it could be anything robotic that you can have sex with. Those devices already exist in the form of sex toys connected to apps that simulate real sensations, for example, or programmable, remote-controlled vibrators. “There are lots of things available right now that are anatomically reminiscent of a human, and facilitate a fantasy experience that is more hands-on than pornography, and more easily controlled than sex with a partner,” says Shelly Ronen, a researcher at NYU who studies relationships, sex, and sex toys.
What people tend to mean when they talk about sex robots, is robots that take a humanoid form
Some of these toys have had success, and some have failed to be marketable. In 2009, a device called the RealTouch hit the market – a device for men that connected to pornography videos and simulated the sensations felt by the actor on the screen in real-time. The experience was quite realistic, according to a review on the technology news website Gizmag. But for all its realism, the RealTouch never found a market, and after a patent infringement lawsuit, in 2013 they stopped sales.
However, most of these devices are sex toys, not robots. What people tend to mean when they talk about sex robots, is robots that take a humanoid form, constructed with the capacity for sexual activity, which have some kind of artificial intelligence allowing them to “think” rather than simply react to sensors.
At the moment, the closest thing to a full humanoid robot is a sex doll sold by companies like the California based company Abyss Creations. Abyss makes and sells something called a Real Doll, a line of realistic human figures which can be fine-tuned with various custom features (right down to freckles on the skin) and “personalities.”
Real Dolls have a dedicated fanbase, but they’re not convincing partners, and they’re not sex robots
Real Dolls have a dedicated fanbase, many of whom connect with one another to bond over the various aspects of a still-unusual relationship. There is even a small group of “doll doctors” who travel around to fix broken Real Dolls. But they’re still extremely expensive – each one will run you anywhere from $5,000 to $10,000 depending on which custom features you opt for. And Real Dolls are dolls, they’re not convincing partners, and they’re not sex robots. At least, not yet.
A true sex robot might track the user’s eyes to see where they’re looking, respond to a user’s facial expressions, and predict or even initiate actions the user enjoys. It would learn the positions and pressure its user enjoys most, ask and respond to questions during the act, and perform the kind of emotional work that sexual partners do.
Sex robots – not dolls, not devices – are a far more complicated prospect. “Sex robots are going to require multiple disciplines to come together, from nanotechnology to replicate non-uniform textures of skin, to an AI complete with natural language understanding,” says AV Flox a journalist who covers the intersection of sex, law and technology. Making a sex robot that really satisfies will require a lot of technical engineering, starting from the skin and going all the way down to the hinges, battery life, and processors inside.
Anyone who has touched a silicone object knows that it doesn’t feel like skin
First, they’ll have to make something that stands up. Right now, the sex dolls and other humanoid forms are heavy (Real Dolls can weigh up to 47kg (105lbs) pounds) and can’t support their own weight. A sex robot will have to be able to not only stand, but also move around on its own and position its own limbs. This isn’t a trivial task. Roboticists are still struggling to replicate smooth human movement.
The robot will also need realistic skin. Anyone who has touched a silicone object knows that it doesn’t feel like skin – and silicone takes a lot of effort to keep clean. Replicating not just the irregularities of human skin, but its feeling, its pliability and give, its stretch and tone, is really difficult. In October of last year, researchers in Singapore announced that they had developed artificial skin that could feel pressure. But it still can’t tell temperature, it can’t stretch, and it doesn’t feel like human skin.
Inside the robot, scientists will have to develop an artificial intelligence that can engage and learn from its human partner. AI has advanced in leaps and bounds recently, but it still can’t simulate much of the emotional labour which goes into sex and relationships. Computers might be able to beat a human at chess, but sex is more like a dance; each partner has to predict and respond quickly to movement. And right now, artificial intelligence and natural language understanding is still a long way from being convincing.
Designers will have to vault the uncanny valley, to create something that is convincing enough to not be creepy
Plus, designers will have to vault the uncanny valley, to create something that is convincing enough to not be creepy. Not even the best animatronic robots can manage that just yet. Madeline Ashby, a futurist and science fiction writer says she thinks early sex robots won’t look fully human. “I think we are more likely to see a cartoony look first, I think that’s how you skirt the uncanny valley problem, is to make a more cartoony or anime-like, or video game-like face and body and appearance.”
These are each interesting, difficult, technical problems. There are people working on all of them individually. But making it happen in reality will require a large team, full of engineers, roboticists, sex toy designers, computer scientists and more to create a convincing sex robot. “They’re not going to be simple enough for a single genius to just put together,” says Flox.
It’s not just the science and research challenges that stand in the way of actual sex robots either. It’s also all the things that will have to align before and after that research gets done – the funding, the laws, the cultural attitudes.
For starters, the arguments for and against these robots are complicated. For example, some worry sex robots might hurt already marginalised sex workers, while others might see the advent of sex robots as a way of keeping those workers safe. The Campaign Against Sex Robots bases much of their argument on the idea that sex work is inherently a bad thing, an idea that has been challenged repeatedly by sex workers themselves as well as organisations like Amnesty International.
Investors are skittish about putting their money behind adult products
There’s also the problem of money. For companies working in the sex industry, investments are hard to come by. In the United States, there are all sorts of both formal and informal rules that make life harder for businesses operating in adult industries. Banks won’t give them small business loans, credit card companies will decline transactions, payment processing services charge them extra fees. Tech platforms like the Apple App store and Google Play won’t approve adult content, whether explicitly adult or simply erotic. Search engines don’t show adult content unless you specifically ask for it, and even then some content is filtered out.
Investors are skittish about putting their money behind adult products, not because they’re prude, “but because they understand that the avenues for bringing these things to market are limited,” Flox says. “The adage that sex sells is true unless you’re actually trying to sell sex.”
And it’s not just in the United States. The second annual Love and Sex with Robots academic conference was supposed to be held in Malaysia in November 2015. But in October, the Inspector-General of Police declared the conference illegal, and they had to cancel it abruptly. “Due to circumstances beyond our control, the Second International Congress on Love and Sex with Robots will be postponed until 2016. The conference will definitely not be held anywhere in Malaysia. We deeply apologise to any person or any authority which have felt offence in any way,” reads the message still up on the conference website.
Sex robots will arrive, in fits and starts and custom runs and highly specialised forms. They will be extremely expensive, and they’ll have to fight regulation and find new forms of income. But the idea that we’re on the cusp of a great sex robot revolution isn’t quite true.
Ronen adds the idea that we’ll suddenly get sex robots, out of nowhere, with no warning, simply isn’t reflective of how technology works. “I think sometimes we imagine that technological advances pop out of nowhere and a commodity plops down in front of us,” she says, “actually technological progress is much more incremental, much slower.
“By the time any sex robot comes, we’ll be so used to having sex with our partners through our computers, that the idea of switching over to something that looks like a partner isn’t going to be so huge of a transition.”
Rose Eveleth is a columnist for BBC Future and the host and producer of the podcast Flash Forward, a show about possible, and not so possible tomorrows. From the completely absurd to the terrifyingly likely, each episode takes on a specific future scenario and tries to really think through how, why, when and if it could ever happen. For more on a future full of sex robots, check out this week's episode.
Join 500,000+ Future fans by liking us on Facebook, or follow us on Twitter, Google+, LinkedIn and Instagram
