In most part of the world people are celebrating Valentine day with their … Valentine. As I am getting ready to discuss with several colleagues at the Future Direction Committee meeting the Symbiotic Autonomous Systems initiative I couldn’t help to wonder if this day could also be a day (in the future of course) for robots to celebrate their Valentine. And this Valentine would be another robot or could it be one … of us?
I am not alone. This question, the more general one of
- Could people fall in love with a robot? or
- Could a robot fall in love (with another robot or with a person)?
has already been asked several times. Movies have touched upon it coming up with stories of robots loving other robots or humans and humans falling in love with robots (I and Caterina was a movie produced in 1980 telling the story of Enrico, a business man, that eventually fell in love with Caterina, a robot), but what about science and technology?
Valentine day is about feeling and emotions, friendship and love. Quite normal words in our human context but really strange in a “robot” context.
Before discussing the robo-context, however, it makes sense to look a little closer to “our” context. Those familiar words have become a focus point for scientists studying the brain, trying to understand the physical underpinning of feelings and emotions. In the electrical-chemical soup in which our billions of neurones and trillions of synapses operate how comes that these words arise and make sense? Just a decade ago it seemed like an impossible quest and indeed for several years the study of the neuronal circuitry from where emotions and feeling arose was set aside. In these last few years, thanks to new technologies that allow scientists to pinpoint the origin of signals and follow them through the brain, more and more data area becoming available and experiments are carried out to verify theories of consciousness, emotions and feelings. A very nice book exploring the latest results in this area is “Synaptic Self: How our brains become who we are” by Joseph LeDoux.
This growing understanding of the hard processes at work in the generation of feelings and emotions is important if we want to answer the above questions on robots from a scientific-technological point of view. Mind you, we are still far from a complete understanding of what is going on in our brain but enormous steps forward have been taken and most scientists agree that it is no longer a matter of “if ever” but of “when” we will have a full grasp of what happens and how feelings and emotions are generated and perceived (i.e. become conscious). First results are already somewhat surprising: we feel, and we experience emotions even before we become conscious of them! This is counterintuitive, since it would seem that first I would have to become conscious of something in order for emotions/feelings to appear. Not so. Experiment have shown, at least for some kinds of emotions that can be tested on animals, like “fear” that the perception, consciousness follows the generation of the emotion.
We already have robots that are -at least to a certain extent- aware of their environment and we have robotic swarms that can become globally aware. This can be a starting point for the generation of fear like feeling and emotion.
We have robots/software that can learn and evolve on the bases of what and how it is learning through experience and self teaching and a few of them have shown a “soul” of their own, a behaviour that surprises their (human) designer.
Clearly, to have a robot engaging in a Valentine day we would need to understand more about this so common and yet so mysterious feeling of love and its related emotion. In principle love as an emotion shouldn’t be so different from fear in terms of its hardware (neuronal) underpinnings, the problem is testing the hypotheses. So far we haven’t seen animals showing anything like love (there are many examples of subsets of love, like caring…) and so testing on animals is not possible, and testing on humans, with the current set of technologies, is out of the question.
Having said that, psychologists and sociologists have started to address the question “can people fall in love with a robot” and surprisingly discovered that, at least in principle, 25% of millennials do not consider unlikely in the future to have a friendship and even a love relation with a robot.
Others go as far as betting that by 2050 artificial intelligence will be so advanced that robots may be undistinguishable from real people, hence sure, we could fall in love with one of them!
On the second question, “could a robot fall in love”, the leading sentiment is that by themselves this is unlikely since robots lack a will of their own but in principle they might be programmed/conditioned to fall in love, eg to show an emotional link to a specific person: imagine you buy a robot from the supermarket shelf and once at home you unpack and power it (him) up you will be the first thing it sees and its program will create a sentimental bind with you (like the ducklings of Konrad Lorentz). A minority of scientists, however, consider that we can only go so far in programming robots and that if we really want them to become intelligent and adaptable to any situations than we will have to let them learn and evolve by themselves. At that point if feelings and emotions (that can emerge from a hardware underpinning) will prove,as they did for us, to be a selective advantage than robots will develop feelings and emotions.
The jury is still out and you can bet what you want knowing that for at least the next 20-30 years Valentine day will be just for us. In the next century it might be different, but it will be something for our grand-grand children to consider.