Why humans prefer robots as flawed as we are

Enlarge Image

Remember those kids in high school who got all A's, never had a bad hair day and never broke a rule but always called you out when you did? They got on your nerves for being so perfect, didn't they? Given that meaningful human interaction often hinges on relatability, we often tend to prefer the company of those who show very real humanness.

The same principle, it appears, goes for robots.

If they have imperfections -- not misplaced screws or actuators, but human-like behavioral flaws -- people are more prone to forge successful working relationships with them, according to a new study out of the UK's University of Lincoln. The findings could have significant implications as people increasingly rely on social robots for tasks like helping seniors stay active and aiding autistic kids in the classroom.

"Our research explores how we can make a robot's interactive behavior more familiar to humans, by introducing imperfections such as judgmental mistakes, wrong assumptions, expressing tiredness or boredom, or getting overexcited," researcher Mriganka Biswas, a Ph.D. candidate at Lincoln's School of Computer Science, said in a statement Tuesday. "By developing these cognitive biases in the robots -- and in turn making them as imperfect as humans -- we have shown that flaws in their 'characters' help humans to understand, relate to and interact with the robots more easily."

The research involved examining the interactions of study participants and two different robots: the ERWIN (Emotional Robot with Intelligent Network), a bot out of the University of Lincoln that can express five basic emotions, and the adorable little yellow Keepon, a robot toy that's been used to study child social development.

During half the interactions the robots acted, well, all-out robotic, but during the rest they showed the kind of "cognitive biases" that make us who we are -- doing things like making mistakes when remembering simple facts and showing extreme happiness or sadness, as expressed through movements and noises.

Participants (the breathing ones with blood flowing through their veins) then rated the interactions.

"We...overwhelmingly found that they paid attention for longer and actually enjoyed the fact that a robot could make common mistakes, forget facts and express more extreme emotions, just as humans can," Biswas said.

Past research has suggested that people feel a more immediate connection to humanoid robots than the boxy industrial-looking sort, though studies also indicate that some people have an Uncanny Valley response to humanoids and believe the ideal social robot shouldn't resemble us too much. The University of Lincoln study takes the inquiry a step further, ascribing human fallibility to our metal friends.

In so doing, the researchers highlight an interesting tension between the superior, distant robots of our sci-fi fantasies and the kind of robots we'd most like to play with our kids and tend to us when we're sick.

"A companion robot needs to be friendly and have the ability to recognize users' emotions and needs, and act accordingly," said Biswas, who presented the study at the International Conference on Intelligent Robots and Systems in Hamburg, Germany, this month.

Next up, Biswas plans to study whether robots that display human flaws and also look like humans could lead to even more human-robot bonding. Watch out, Little Miss Perfect Robot, your days might be numbered.