The Uncanny Valley of Personality: People Like Robots Better When They Aren't Perfect
According to the theory of the "uncanny valley," people are viscerally repulsed by artificial beings that look somewhat like humans, but are too plastic and perfect. Now, researchers from the University of Lincoln have procured evidence that there exists an uncanny valley of personality, in which humans prefer robots who are at least slightly flawed in their functioning.
For the experiment, members of the public were asked to interact with two different robots: the Emotional Robot With Intelligent Network (ERWIN), which was created in order to explore the possibility of long-term relationships between humans and robots and can express five different emotions in its facial expressions, and Keepon, a small yellow robot who resembles a rubber ducky and was created to study social development by interacting with children. For the first half of the experiment, both robots carried out their tasks impeccably, but in the second half, ERWIN experienced memory glitches, and Keepon exhibited extreme swings between happiness and sadness. The researchers found that participants overwhelmingly preferred both robots when they made mistakes.
This makes sense, considering that humans experience forgetfulness, mood swings, and general imperfections all the time. Humans are fallible, so it makes sense that we would more easily relate to robots who are also imperfect. Plus, they're infinitely more entertaining when they fail at tasks than when they succeed:
The researchers hope that this research could lead to more effective companion robots with "personalities" that more closely approximate human behavior, which could be used in the future for tasks like elderly or childcare (because that always goes so well):
The idea of giving robots personalities is certainly not a new one; the trope is nearly ubiquitous in science fiction, and there's a chance that we want endearing robots who make flubs because they remind us of Wall-E. But, according to the researchers, this is likely a very innate reaction, similar to the uncanny valley, and as a result it is a problem that needs to be solved in order to create effective companion robots.