Near-human robots: easy to trust but hard to love

You may be looking forward to the day when your robot slave does all of your bidding. But even so, you probably don’t want to live with one of those weirdly too-human looking robots. At least, that’s the finding of new research published this week.

Researchers have long known about this paradox, according to a story in New York Magazine. As robots become more of a part of daily life, designers look to build ever more human-looking robots. Surprisingly, surveys have found that we are comfortable around cute, loveable robots like Wall-E, but as robots get to looking closer to human, there is a point at which people find them less likeable, and even creepy or threatening.

This effect, known as the “uncanny valley” concept, is the subject of new research published in the journal Cognition by Stanford University social scientists Maya B. Mathur and David B. Reichling. Their surveys found a difference in human perceptions of robots when it comes to trust, versus likeability.

Their work finds that, as robots appear more human, people are more likely to retain their trust of the robots, even as the robots become less likeable. As robots are likely to become more prominent in many people’s daily lives, the researchers say that designers need to understand and account for these perceptions during robotic development.