Unusual relationships

From the American woman who married the Eiffel tower to the Japanese man who is happily married to his favourite Nintendo character. There is also a small, but growing number of men who are in a relationship with a sex doll (or, as the company that makes them puts it: a love doll). Some people even claim that the sex industry will become the driving force behind the rapid developments in artificial intelligence aimed at equipping sex dolls with complex artificial intelligence (AI) so they can also make interesting conversation*. Quite apart from the fascination of the technology involved, such developments could also help to solve societal problems.

An enduring romantic relationship with a robot?

The Swedish TV-series Real Humans, for example, affords us a glimpse of a future in which robotics create a socially acceptable form of prostitution. But given the rapid advances in AI, having sex with a robot is just the tip of the iceberg. Long-distance relationships between people who have never met are not unusual (see the wonderful TV-programme Catfish), and this situation may persist for months, or even years. So good conversation seems to be pretty important for relationships. Once AI finally passes the Turing test, will we perhaps be able to have an enduring romantic relationship with a robot? And why might that be desirable? I believe that in the long term robots could provide a solution for people who want a romantic relationship but, for whatever reason, cannot find a partner. That would be good news, because research shows that being single is far from good for your health. A meta-analysis shows that single men die between eight and seventeen years earlier than married men, and single women between seven and fifteen years earlier than married women (Roelfs et al., 2011). Of course it remains to be seen whether robots can make a difference to these statistics, but it is not unthinkable that they might play a role in combating loneliness. Only time can tell what the next major step will be in AI, but we can make an educated guess on the basis of current developments.

What makes us ‘human’?

Super computers such as IBM’s Watson provide us with phenomenal processing power. When this power is harnessed to clever algorithms, such computers can beat humans at chess, develop new drugs, and even win at Jeopardy. But cognitive scientists are increasingly convinced that what makes us ‘human’ is not our brain’s huge processing power, but its architecture. Relatively simple elements, neurons, with many interconnections and operating in parallel with one another are apparently the secret behind human beings’ rich experiences and complex behaviours.

An accurate simulation of a human brain

In conceptual form, this idea is already being used in a whole range of AI applications, but the Blue Brain Project, headed by the École Polytechnique Fédérale de Lausanne, takes things one step further. In the quest for true artificial intelligence, they aim to produce an accurate simulation of a human brain at the neuronal level. After all, if human behaviour derives from the human brain, then surely a digital copy of this brain should be just as clever, funny, and sensitive as we ourselves?

Free will is a double-edged sword

Whether or not this is actually the case, the whole endeavour gives rise to a whole series of questions. What if at some point we are in a position to run a simulation of this kind in a true-to-life robot? It seems more than likely that a robot of this kind would elicit in us (and experience!) exactly the same emotions – and maybe even romantic feelings – as a human being. So we also need to pause and reflect on the ethical dilemmas that this conclusion entails. If we can attribute to a robotic partner the same experience of emotions (and all that goes with it), the advantages that a robot partner may seem at first sight to possess may well not be borne out. Free will is a double-edged sword, after all. How could we defend wanting a romantic relationship with a given robot, with its own will and feelings, if it would actually prefer a different partner. People who turn to robots to fill the void left by the absence of human contact may perhaps end up discovering that robots are also not all that attracted to them either.

A sort of philosophical zombies?

As soon as we start to entertain doubts about robots’ inner experience, we must also think critically about the role we want robots to fulfil in our society. If robots are no longer our slaves, but have a full role in society, what use do they serve other than that of a tour de force of cognitive science? If we want only to use social robots – for romantic or sexual relationships, or in other contexts – it is important that they should behave like humans without experiencing the associated feelings; they should be a sort of philosophical zombies.

Research into Human Robot Interaction

Whether it comes about through simulating the brain at the neuronal level, or through symbolic AI, the fact remains that robots increasingly resemble humans, with more complex behaviour and human-like bodies. To prepare ourselves for a society in which robots and humans work together on a large scale, we need a strong field of research into Human Robot Interaction (HRI), that can join the cognitive sciences in ensuring that this development doesn’t get out of hand. If we don’t keep things in check, we shouldn’t complain if in twenty years’ time old people’s homes are full of neglected and depressed robots. But hey, that’ll mean more work for psychologists, I guess.

*Perhaps somewhat similar to the myth that the porn industry was the moving force behind the war between VHS and Betamax in the 1980s.