SEATTLE -- Last year, when Steven Spielberg brought out his tech version of Pinocchio, AI, the question he raised was whether a robot could ever really learn to love. Now, husband-and-wife researchers at the University of Washington are asking whether young children will learn to love a robotic dog, and what effect will it have on society if they do. In a demonstration for the press on the Washington campus Wednesday, a group of 3- to 5-year-olds played with a Sony Aibo robot dog, and a soft plush black-lab puppy while answering questions about how they saw the two. Ann Foreman, a senior in the Information School asked the children if they thought the toys were alive, whether they would get hurt if she dropped them on the floor, and what they would do if "Aibo" or "Shanti" got broken.
The kids agreed that neither the metallic nor the stuffed dog could feel pain because they were toys. All but one of the children said that Aibo was not alive. (One girl said it was "because she moves.") Their suggestions for what to do with a broken Aibo were to throw it away or recycle its parts. At the same time, they said that Aibo did have a stomach, and they held and petted it carefully, the same as if it were a real dog.
"One of the ways in which children learn to be responsible for others, that their actions can affect others, that there can be negative consequences, is through their interactions with animals," said Batya Friedman, an associate professor at UW's Information School. "As they develop those senses, then they carry those over to their interactions with other people."
Programs already exist for cars that can converse to keep the driver awake, and similar devices are on the way for everything from thermostats to toasters. No one is expecting people to begin mistaking their refrigerators for friends.
Friedman said she is not concerned that a visit with a robotic dog, like the Sony Aibo, will be harmful. But she worries that if parents decide to substitute robots for the real thing, kids may miss out on the lessons they learn from having to care for something that needs to be fed and cared for to keep it alive and well.
Friedman and her husband Peter Kahn, a UW psychology professor, are examining the way preschoolers perceive the robot dogs as part of a wider study with a research team at Purdue University looking at how people respond to a variety of simulated versions of reality.
The Purdue researchers have examined the effect Aibo has on nursing home residents, who cannot have real pets. They found that the elderly get many of the same benefits, from simple pleasure to lowered blood pressure, by holding and caring for the mechanical dogs.
"I think the movements of the robotic dog are pulling at us at a level below cognition," Friedman says in trying to explain those results. "Rationally, the kids don't think this is alive, the elderly don't think this is alive. But we've got enough cues to things that we respond to in a social way, that whole framework that we have comes into play psychologically and we get some benefits from that."
Adds Kahn: "We are hopeful there will be real benefits for the elderly, who may no longer be capable of caring for real animals.
"However, with children in those early stages of development we have concerns about what happens when they fall prey to accepting robotic companionship without the benefits that real companionship involves," Kahn said.
"The main point I see is that if these robotic interactions were simply in addition to rich interaction with nature and animals, this is great stuff. It's bringing kids into the technological world. The issue, I think, is they're going to be replacing, to some extent, interactions with live dogs. Once it becomes replacement, I think you have problems on the moral dimensions."
The two teams are conducting other virtual reality experiments as well. In one case, they're comparing the effects of gardening by remote control with actual hands-in-the-earth gardening. In another, they're examining the effects of looking at a high-definition screen connected to a camera to the outside, versus having a real window looking out at the same scene.
In both cases, Friedman and Kahn said they haven't finished analyzing their data and aren't ready to publicize any conclusions.
Aibo Owners Biting Mad at Sony
New Pet Robot Won't Stain the Rug
