UW News

May 13, 2002

Researchers studying whether a robot dog is kid’s best friend

Pets can help children learn about life, love and death. University of Washington researchers are studying whether today’s robot pets can do those things, too, or whether they fall short when it comes to stimulating a preschooler’s moral and intellectual growth.

The researchers — a psychologist and a computer and information scientist — say their findings could affect how society views not just sophisticated interactive dogs but also the growing array of other chatty “smart” devices in homes, offices and cars.

In the study, 80 preschoolers spend 40 minutes apiece with AIBO — Sony’s $1,000 metal-skinned canine — while researchers ask questions and observe.

The youngsters (ages 3 through 5) cuddle and “train” the amazingly frisky AIBO much as they would a real dog. Missing, say the researchers, are the feedback and consequences that come with handling a live animal.

“It’s not a real, reciprocal relationship,” said Peter Kahn, a UW research associate professor of psychology. “You don’t have to feed AIBO, and he doesn’t really care about you, even if it seems like he does.”

Co-researcher Batya Friedman, an associate professor in the UW’s Information School, said AIBO and its brethren have gotten increasingly lifelike, thanks to software and sensors that allow them to “learn” and interact. Even young children know perfectly well that AIBO is not a live pooch, but its realistic responses and simulated emotions make it a very different animal from, say, a plush toy.

To understand how AIBO differs from stuffed dogs, the researchers record what children say to each and, more importantly, what they do. Kids in the study often cuddle stuffed animals, but they also toss, squish and sit on them. With AIBO, on the other hand, the tots tend to touch the robot gently, pet it behind the ears and stroke its back — similar to the way children treat a real dog.

“With a stuffed animal, children tend to have a rich fantasy life,” said Friedman. “With AIBO, it could be confusing the boundaries between what’s real and imaginary because you get clues that prompt a real social rapport.”

The research is funded by the UW Center for Mind, Brain & Learning.

Friedman and Kahn (with colleagues at Purdue University) also are putting AIBO through its paces at the other end of the age spectrum — with nursing home residents, in a study to see whether robotic pets could provide some of the benefits of live ones without the worries.

“We are hopeful there will be real benefits for the elderly, who may no longer be capable of caring for real animals,” Kahn said. “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 developmental benefits that real companionship involves.”

These AIBO studies are part of a larger research program by the two scholars funded by the National Science Foundation to understand the psychological effect of technology that augments the human experience of the world around them.

“In the coming years,” Friedman said, “robotic pets will become more technologically sophisticated — more animal-like. As they do, our research suggests that they will evoke more and more psychological responses from humans. Is that a good thing?”

BACKGROUNDER: Exploring preschooler-robotic pet interaction is part of larger UW effort to study psychological effects of information technology

As sensors, chips and sophisticated software become increasingly ubiquitous in everything from toys to talking toasters, little thought and even less scientific research has been devoted to the psychological and ethical impacts of such devices.

That’s changing. Today’s demonstration of how children interact with a robotic dog shows that these devices evoke many human-animal-like interactions, such as petting, caressing and playing fetch. But robotic dogs don’t need to be fed or taken on walks, and you can stick them in a closet when you decide to go skiing for the weekend. Thus long-term interaction with robotic dogs may diminish people’s moral relationships with animals — and even with other people.

Today’s demonstration is just one element of a broader University of Washington research program. This program seeks to show how parents, educators, consumers, and scientists — indeed society at large — can shape information technology to better support human lives. The chief investigators are Peter Kahn, UW associate research professor of psychology, and Batya Friedman, associate professor in the UW’s Information School. In addition to the preschooler-robot dog study, the program is currently investigating:

***Room with a view. Just looking at nature can reduce stress, and virtually every worker yearns for a window to look out of. This experiment uses high-definition television to project a real-time image of UW’s Drumheller Fountain onto a large video plasma display in an office.

The UW researchers seek to understand how such projection technology can be used to foster physiological health and productivity in the work place, creativity, and a sense of well-being. They are also investigating the ethical issue of privacy for people walking by the fountain.

*** Growing green thumbs. Subjects in this study will travel on-line to a telegarden where they can remotely “garden” by controlling a robot in a distant but real garden. Physiological and observational data will be collected. Subjects will also be questioned about their perceptions, values and emotional responses to the experience.

*** AIBO and older kids. The interactions of three groups of older children — ranging in age from 7 to 14 — with a live pet and AIBO are currently being examined in this joint investigation with Alan Beck and Gail Melson at Purdue University. As robotic pet technology becomes increasingly sophisticated, the researchers are interested in studying what design features future robot dogs should have.

***AIBO and the elderly. Research already has shown that animal contact contributes to the physiological and emotional well-being of older people.

However, illness, physical condition and housing arrangements often prevent older people from having pets. The UW team, in collaboration with Alan Beck and Nancy Edwards at Purdue University, is investigating whether robotic pets such as AIBO can provide the elderly with some of the benefits that come from playing with and caring for real pets.

*** Going on-line with AIBO. In this completed and published study, Kahn, Friedman and UW undergraduate student Jennifer Hagman, analyzed people’s concepts about AIBO from postings on three online discussion forums devoted to the robotic pets.

Few of the participants engaged with AIBO in moral terms. People didn’t believe, for example, that AIBO deserves respect, has rights or can be held morally responsible for its actions. However, the majority of people did say their robotic dog has physical properties such as a brain, that its actions could be attributed to different states of mind, and that their AIBO provides companionship. One AIBO enthusiast wrote, for example: “I feel I care about him as a pal, not as a cool piece of technology.”

The researchers are concerned because people in general, and children in particular, may fall prey to accepting robotic companionship without the moral responsibilities involved in reciprocal companionship. But they are hopeful that for certain populations, such as the elderly who might not be capable of caring for real animals, robotic pets may offer some degree of companionship and health benefits.

The AIBO preschool study is being funded by the UW’s Center for Mind, Brain & Learning. All of the other above studies are being funded by the National Science Foundation.

For more information, contact Friedman at (206) 616-2548 or b