Henry Wellman is the Harold W. Stevenson Collegiate Professor of Psychology at the University of Michigan. Kimberly Brink is a doctoral candidate in developmental psychology at the University of Michigan.
Science fiction writer Isaac Asimov collected a series of his short stories on robots in his now famous anthology I, Robot.
The series "revolutionized science fiction ... and made robots far more interesting than they ever had been," according to the Saturday Evening Post.
I, Robot begins with a lesser-known story: Robbie. Robbie is an experimental robot brought home by George Weston, a robotics engineer, to nanny his 4-year-old daughter Gloria. Quickly, Gloria and Robbie become inseparable. He plays games with Gloria, and she speaks for him and tells stories to the mute quasi-humanoid device. Gloria tells her parents that Robbie is her best friend, she holds his metal hand, and she shares secrets and tears with him.
Meanwhile, Grace Weston, Gloria's mom, goes from uneasy to increasingly concerned about Robbie and her daughter. "It's Gloria and that terrible machine...You listen to me, George. I won't have my daughter entrusted to a machine — and I don't care how clever it is. It has no soul, and no one knows what it may be thinking."
This is fiction — from a time when there were no actual humanoid robots. Now there are. You can find robots in malls, hotels, assembly lines, hospitals and, of course, research labs.
The National Robotics Initiative foresees a future in which "robots are as commonplace as today's automobiles, computers, and cell phones. Robots will be found in homes, offices, hospitals, factories, farms, and mines; and in the air, on land, under water, and in space.
And just as Asimov foresaw, some robots certainly make adults uneasy. Decades of research reveal that while adults prefer robots that are somewhat human-like, they find very human-like robots unnerving. This is known as the Uncanny Valley. Both by hypothesis and according to research, machines become increasingly attractive as they become more human-like until they reach a threshold at which they become too human-like and are considered creepy.
It's this precipitous dip in affinity for very human-like robots that is the Uncanny Valley. Very human-like robots are distinctly creepier than other robots and, in particular, creepier than even unsettling machine-like robots (e.g., clearly machine-like robots whose metal gears and wires are more obvious).
Some scientists think very human-like robots are creepy because we evolved a fear of illness and very human-like robots tend to look like sick humans. Others argue that closely human-like robots give the impression that they can think and feel, but we, as adults, don't believe machines should be able to think or feel. So, robots that look and act like they can violate our expectations are unsettling.
Either of these factors could have caused Grace Weston's fictional uneasiness. Given that Robbie was reasonably human-like, Asimov's story could well have been prophetic. His portrayal of Grace was close to the mark. How about his portrayal of Gloria?
While there were many studies on the Uncanny Valley with adults, there had been no research with children. Do children experience the Uncanny Valley, too?
The payoff from knowing how children feel about robots would be far from merely theoretical: Robots are already entering homes, not only to help adults with household chores, but also to play with, teach, and tutor children. Dozens of robots have been released in the past year alone designed specifically to interact with children. So, how do children think and feel about robots?
So, we interviewed 240 children (3- to 18-years-old) — alas none named Gloria — about their beliefs about the minds of three different robots and how they felt about these robots. Specifically, we showed them videos of a very human-like robot, a machine-like robot, and one humanoid robot that looked like a combination of the fictional robots Baymax from Big Hero 6 and EVE from WALL-E. We asked the children if they thought the robots could think for themselves, whether they could do things on purpose, and whether they knew the difference between right and wrong. We also asked if they thought the robots could feel hungry if they didn't eat breakfast, feel scared if they saw a snake, or feel pain if we pinched them.
And focally, of course, we asked how creepy the robots were and if they made the children feel weird.
Young children (children younger than 9 years), unlike adults, don't think robots, even those that look very much like humans, are very creepy at all. So, it's unlikely that the prime reason that human-like robots are creepy for adults is due to an evolved aversion to sick humans. This would have meant that even very young children should find human-like robots creepy. They don't. In real life, like in Asimov's fantasy, Gloria doesn't feel the way Grace does.
Older children in our study (children older than 9 years), on the other hand, did find the human-like robot much creepier than the machine-like robot. This shows that our negative reaction to human-like robots is likely something that is learned over development.
So what does this mean about the Uncanny Valley? One feature of our results provides a clue: Children's attributions of mind to the robots affect how children feel. The younger children preferred a robot when they believed the robot could think and make decisions. For them, the more mind the better. This is in contrast to adults and older children for whom the more robots seemed to have minds (and especially minds that could produce and house human-like feelings and thoughts) the creepier that made them. For adults and older children, a machine-like mind is fine, but a robot with a human-like one is out of bounds. Perceived creepiness is related to a perceived mind. This finding also highlights the importance of thinking developmentally.
Children are not little adults. They have different ideas about the world that change with age and experience. Eight-year-old Gloria and her peers are happy to believe robots can do all kinds of things, like see, hear, think, play and cry. And when they do these things, that seems to make robots comfortable and familiar. And so, such robots can potentially be their friends and even protect them. For them, more mind makes all that seem more likely.
So, Asimov's portrait of Gloria, who turns 8 as the story unfolds, and whose character and reactions drive the whole story, was not so far fetched. Fiction, but not pure fantasy.
Robots are becoming more and more present in the lives of children. They are not just designed for homes but are already teaching children in classrooms and helping them in hospitals. Our findings highlight the importance of considering the appearance of robots that will interact with children.
And this means thinking developmentally, thinking carefully about children of different ages. For young children, it might be helpful to design robots so that they appear to have minds of their own or look more human-like. For older children, however, we might want to avoid designing robots that look more like humans.
Thinking developmentally means thinking carefully about when children were born, too. Children born during the Depression grew up differently than those born in World War II, Baby Boomers, and Millennials. In the future, children who grow up with humanoid robots as a constant fixture of their lives, like Gloria with Robbie, may never come to think of closely humanoid robots as creepy. Future adults may outgrow today's Uncanny Valley. Hi, Robot.