Robots, according to a study released this week, are able to manipulate humans.
Researchers from the University of Duisburg-Essen in Germany studied how humans interact with robots. In a paper published Tuesday in PLOS ONE, they described a series of experiments in which robots attempted to stop humans from switching them off.
The team used 89 volunteers who interacted with the robot—a small humanoid called Nao. The volunteers were told they were interacting with the robot to help it become more intelligent.
At the end of their interaction, a researcher asked the participants to turn off the robot. But the robot then begged the participant not to, using bodily actions to reinforce the point. (A control group were asked without the begging.)
Of the volunteers, 43 were begged not to turn the robot off. About 30 percent, or 13 participants, followed the robot's request instead of the researcher's. Compared to the control group, each of the 30 other participants took longer to turn off the robot. The type and length of the interactions between the robot and volunteers didn't have any impact on the response of the person when they were asked by the robot not to switch it off. The scientists think this means that humans can view humanlike robots as autonomous machines. Which means that robots have the ability to manipulate humans.
When the volunteers were asked why they refused to turn the robot off, some said they were afraid of doing something wrong or they felt sorry for the robot. Many said they didn't switch the robot off just because it asked them not to.
The scientists also tested which voices were most effective in convincing people not to turn the robot off and found that a fearful protest was the strongest. But people hesitated the longest when the robot expressed a wish to stay turned on.
The team also found that technophiles hesitated the longest, likely because they would be fascinated to hear what robot would say next, and those who had negative opinions of robots would switch it off without hesitation.
