Humans like error-prone robots more than faultless robots. That according to new research published in Frontiers in Robotics and AI.
"We conducted a user study for which we purposefully programmed faulty behavior into a robot's routine," explains Nicole Mirnig, lead author on the paper. "It was our aim to explore if participants rate the faulty robot different from an error-free robot and which reactions people show in interaction with a faulty robot."
For the study, human participants listened to identical NAO units--the sometimes cheeky humanoid from Softbank--ask a series of questions and then give directions on completing a task involving Legos. Some of the robots did the job flawlessly, but in other cases the researchers introduced a series of slip-ups.
"Our results show that participants liked the faulty robot significantly better than the robot that interacted flawlessly," writes Mirnig. Interestingly, there were no significant differences in people's ratings of the robot's anthropomorphism or perceived intelligence--they just plain liked the error-prone bot more.
That may not be a huge surprise to fans of sci-fi. The neurotic C-3PO, if not exactly faulty, elicits our grinning sympathy because of the sharper edges of his fastidious personality. At the risk of extrapolating from the findings, it would seem we feel more positively toward things that seem more imperfect and human in their actions.
The researchers didn't measure how flawed a robot can be before likability begins decreasing, though presumably there's a limit.
The findings are significant for the field of social robotics. With advances in interactive AI and voice recognition, and with widespread consumer adoption of personal assistants like Alexa, Siri, and Google, the pump is primed for a new generation of robots in the home.
When your next droid strategically drops a hamper full of laundry, you'll know who to blame.