This post originally appeared on IEEE Spectrum’s Automaton blog.
Last year, we posted about a study investigating whether people care if a robot friend of theirs gets unfairly stuffed into a closet, featuring one of the saddest robot videos ever. Turns out, people do care. A lot. And they care even though robots don’t have feelings. Now, we’re looking at another study from researchers at the University of Duisburg-Essen in Germany that uses a functional MRI procedure to see just exactly how much people empathize with robots compared to humans.
IEEE Spectrum contributor Charles Q. Choi has a detailed story on this, with psychologists and experts in human-machine interaction describing how they see the experiment. Below, I include some more details and also my own take on the results.
From the study:
In the first study, 40 participants watched videos of a small dinosaur-shaped robot that was treated in an affectionate or a violent way and measured their level of physiological arousal and asked for their emotional state directly after the videos. Participants reported to feel more negative watching the robot being abused and showed higher arousal during the negative video.
The second study conducted in collaboration with the Erwin L. Hahn Institute for Magnetic Resonance Imaging in Essen, used functional magnetic-resonance imaging (fMRI), to investigate potential brain correlations of human-robot interaction in contrast to human-human interaction. The 14 participants were presented videos showing a human, a robot and an inanimate object, again being treated in either an affectionate or in a violent way. Affectionate interaction towards both, the robot and the human, resulted in similar neural activation patterns in classic limbic structures, indicating that they elicit similar emotional reactions. However, when comparing only the videos showing abusive behavior differences in neural activity suggested that participants show more negative empathetic concern for the human in the abuse condition.
Here’s one of the videos used in the study. Warning: not safe for lovers of robots.
OK, well, I suppose it’s probably unrealistic to expect that humans should feel just as empathetic toward robots as they do toward other humans, but what I’m really curious about at this point is how much more empathetic we (might) feel towards an animal than towards a robot. And you’ve got to figure that there’s a line somewhere. Like, would we feel more empathetic toward dogs and cats than toward robots? Sure. Rabbits? Probably. Turtles? Maybe. Insects? Probably not. But where is that line, and why does it exist where it does? And what does that say about us, and the future of human-robot interaction?
Here’s what really kerfuffles my noggin: When I read about these sorts of studies, I think to myself, “I know exactly what’s going on here.” I mean, I know that this robot is getting a certain pattern of sensor input from the user, and that input pattern is calling up some code written by a human that instructs servos to move and sound files to play. The robot isn’t feeling anything. It’s just executing a series of commands that are specifically designed to exploit certain human emotions.
But it doesn’t matter.
I have a Pleo. I would never do this to my Pleo. Never, ever, ever. Because it would be cruel. I don’t care that it’s a robot, and I don’t care that it doesn’t have emotions, because I have emotions. And ultimately, that’s what it’s about, isn’t it? So maybe this research just says good things about humans. Maybe this means that we really are decent folk, and that we’ll treat robots well because not treating robots well reflects badly on ourselves.