Star Wars’ R2-D2 shows that a robot—even one that looks more like a trash can than a person—can make people laugh and cry. Now, in research to be presented at the International Communication Association conference in London, scientists have shown that when the human brain witnesses love for or violence against a robot, it reacts in much the same way as if the robot were human.
Engineers worldwide are developing robots to act as companions for people—for instance, to help the elderly at home or patients in hospitals. However, after the novelty of using a robot fades, people often feel less interested in using them. Scientists want to learn how to create more-engaging robots, but there has been little systematic research on how people react emotionally toward them.
To learn more, social psychologist Astrid Rosenthal-von der Pütten, at the University of Duisburg-Essen, in Germany, and her colleagues had 14 volunteers watch videos as they scanned their brains with functional magnetic resonance imaging (fMRI). In some of the clips, experimenters treated either a woman in a green T-shirt, a green box, or a small green dinosaur-shape robot with affection—for instance, with tickles, hugs, caresses, or massages. The robot, known as Pleo, had microphones, speakers, and motors, as well as light, infrared, motion, touch, and grip sensors. When it was stroked or “fed” with items placed in its mouth, the droid purred, sang, babbled, and squealed with glee.
In other videos, experimenters acted violently toward the targets—for instance, strangling them with a rope. When the robot was struck, choked, shaken, beaten, put into a plastic bag, or dropped, the robot behaved as if it were suffering by crying, bawling, rattling its breath, and making choking and coughing sounds; the woman made comparable sounds. The idea of torturing a robot on camera was inspired by a video that recorded how much punishment a robot dinosaur could take. “I was amused about the idea of the video, and at the same time I felt kind of bad for the little dinosaur robot,” Rosenthal-von der Pütten recalls.
The researchers were surprised by how similar the brain responses to both humans and robots were. “Even though we assumed that the robot stimuli would trigger emotional processing, we expected these processes to be considerably weaker than for human stimuli,” Rosenthal-von der Pütten says.
Affectionate behavior toward the robot and the human led to similar levels of brain activity in limbic structures of the brain, which are regions linked to emotions. Violent behavior led to similar brain activity as well, although differences suggest the volunteers showed more concern for the human than for the robot.
“It is not surprising that we have empathy to robots. Everybody understands that they might react strongly to R2-D2 being trashed with a sledgehammer,” says psychologist Arvid Kappas, at Jacobs University Bremen, in Germany, who did not participate in this study. “The present study is one of a group of emerging studies that tries to use responses in the brain and the rest of the body to better understand who bonds how and why with machines....This is a great advance.”
This research shows “we are vulnerable to seeing robots are appropriate for things they are not appropriate for,” says psychologist Sherry Turkle, at MIT. Turkle, who has investigated human-machine interactions extensively, has long argued that humans can form unhealthy relationships with machines.
“They are not who we should be talking to in things regarding human meaning,” she says. “The fact that you can be fooled into talking to a robot—this is something that we need to be vigilant about. It is not an opportunity to make robots more like people. It means we have a responsibility to make them less like people.”
Such research could help make robots more or less sympathetic as needed, Rosenthal-von der Pütten says. This work is not aimed “to replace humans or isolate them but rather on facilitating interactions with technologies,” she cautions. “We think that technology has great potential to support many people in diverse tasks and situations. Our work concentrates on making interactions as easy and natural as possible.”
Further research is needed to discern the factors involved in long-term reactions to robots. “In previous research, we observed that during a long-term interaction with an assistive robot, some participants showed signs of bonding and relationship building with the robot, while others treated it like a piece of technology,” Rosenthal-von der Pütten says.
Scientists could investigate how people respond to other kinds of robots, including humanoids, says roboticist Luke Wood, at the University of Hertfordshire, in England. “A humanoid robot may have produced even more similar neural activation patterns,” Wood says. Future studies could also explore the reactions to verbal and other kinds of abuse against such robots, Rosenthal-von der Pütten adds.
Such research might also shed light on empathy in general. “Atrocities toward others, particularly in wars or violent conflict, but also in more normal situations toward outsiders, suggest that we tend to see some humans as less human,” Kappas says. “In turn, some people treat completely inanimate objects, such as a car, as if they had a soul. What we are seeing is that there are likely spectra of ‘humanness’ or ‘thingness’ that overlap.”
In the long run, such research may “perhaps be good for public discourse on how we want to deal with sentient machines when they are ready,” Kappas says.
About the Author
Charles Q. Choi is a science writer based in New York City. His reporting has appeared in The New York Times, Scientific American, and Wired, among other publications. In December 2012, he reported for IEEE Spectrum on a hack that targets Cisco IP telephones.