In movies, robots typically masquerade as humans in order to infiltrate society and annihilate us. In real life, researchers are more concerned about the damage androids could inflict on our minds. Alan Winfield, a roboticist at the Bristol Robotics Laboratory in the U.K., says we need rules for how lifelike a robot can appear–even those built to work as companions or caregivers. “It’s unethical to build a robot that looks like a human but is not much smarter than a washing machine,” he says. “It’s a deception.”
Humans are “pathological anthropomorphizers,” Winfield says. And that puts us at risk of becoming emotionally attached to a machine that can’t reciprocate. It’s the sci-fi equivalent of a one-sided friendship (or, possibly, robomance) with all of the corresponding social costs. When Winfield and his colleagues drafted ethical guidelines for creating robots in 2010, they included a restriction on deceptive systems. A robot’s appearance, they wrote, should expose its mechanical form. “You should always be able to pull the curtain aside to reveal the machine,” says Winfield, “just like Toto did in The Wizard of Oz.”
Others are not so sure. Karl MacDorman, an Indiana University roboticist who specializes in human-robot interaction, says robotic companionship can be therapeutic. “With elderly people, the biggest concern is the three D’s–dementia, delirium, and depression,” he says. Robotic pets have been shown to elevate mood and decrease stress in the elderly in Japan; an android could likewise reduce the social isolation that’s believed to exacerbate the three D’s. MacDorman agrees that humanlike bots could interfere with our social instincts. But if the alternative to a fake friend is debilitating loneliness, perhaps a dose of deception is just what tomorrow’s doctors should prescribe.