It’s a medical device that looks like a baby harp seal, and someday it just might be your best friend.
Which leads to some interesting questions. Like, Why would you want to hang out with a medical device? And, Is this a good idea?
The device is called Paro. It’s cute and cuddly, with soft white fur, big bright eyes and — awww! — a pacifier to suck on. Casual observers could easily mistake it for a typical stuffed animal toy.
But Paro is no plaything. It’s a robot specially designed to provide comfort and companionship to the elderly.
Evidence to date suggests that it works. This raises the hopes of many people concerned about the best way to care for a vulnerable population.
And raises the hackles of others.
The elderly have serious needs that often go unmet, says Maja Mataric, co-director of the USC Robotics Research Lab. “Robots aren’t a panacea, but if they help people, if they can make their lives better, then what’s wrong with that?”
What’s wrong, others believe, is that robots are inherently deceptive.
“Robot companions are not companions,” says MIT’s Sherry Turkle, who studies the ways people interact with technology. “They are not competent to be companions.”
Opinions on both sides converge on one point: Rapidly advancing robot technology presents challenging ethical issues for the healthcare community. At heart, it’s a question of how the technology is used, says Dr. Tia Powell, director of the Montefiore-Einstein Center for Bioethics in the Bronx. “Every human tool can be used wisely or poorly.”
Why make a robot that looks like a baby harp seal? Takanori Shibata, a research scientist in Japan’s Ministry of Economy, Trade and Industry, was going for adorable when he designed Paro in 2003. And very young harp seals are definitely that, with spectacular snow-white coats to go along with irresistibly winsome faces. Mystery solved.
(Paro also sounds like a baby harp seal. Some find the strange squeals a baby harp seal makes quite resistible indeed, making this design choice rather less obvious.)
What can robots do?
Robots aren’t just about being smart anymore. Paro, which was designated a Class II medical device by the U.S. Food and Drug Administration in 2009, is only one of a growing category of robots classified as “social” because they’re designed to interact with people instead of, say, trouncing them at Jeopardy.
Paro, available for $6,000 from Paro Robots U.S. Inc. in Itasca, Ill., is built to behave as though it’s a real animal — maybe not a real baby harp seal, but a real baby of some species that lives indoors and craves being held and fussed over by humans. You can name it whatever you want, and it will “learn” that that’s its name. It will also respond to how you treat it, acting happy when you’re nice to it and unhappy when you’re not. If you pet it, it will remember what it was doing beforehand and do it again (and again). But if you punch it or yank its tail (what kind of monster are you?), it will avoid doing whatever seems to have prompted that punishment.
These sorts of people skills are intended to make Paro a good companion for elderly patients in hospitals and extended care facilities, helping them feel less lonely, isolated, anxious and agitated.
Just ask Eve Anthony, chief operating officer of the Athens Community Council on Aging in Athens, Ga. The council uses a Paro named Bentley with individuals and groups in its adult day health center.
One client there is often so agitated that the staff has a hard time taking her blood pressure or even getting her to eat. “But we’ll just hand Bentley to her, and it really calms her down,” Anthony says. “She’ll sit with him a long time and just pet him and love on him.”
It takes data to really assess how much help robots can provide, of course, and a number of small-scale studies have made attempts at figuring this out. For example, a 2005 study published in Proceedings of the IEEE International Conference on Robotics and Automation found that residents of a nursing home spent more time in the common areas and socialized more if a Paro was there too. Urine tests of the residents showed that their stress proteins went down after being around the little seal.
In another study, published in 2009 in the Journal of Advanced Computational Intelligence and Intelligent Informatics, nursing home residents interacted with Paros more than nine hours a day for a year. Afterward, researchers found that residents communicated more and had stronger ties to each other.
A 2008 study in the Journal of the American Medical Directors Assn. tested how nursing home residents responded to another companion robot, Sony Corp.'s canine-styled Aibo (no longer in production). Some residents got a weekly 30-minute visit with Aibo while another group got the same amount of time with an actual dog named Sparky and a third group didn’t spend time with any sort of four-legged creature. After seven weeks, the residents who had been hanging out with Aibo or Sparky were less lonely than those in the control group. Moreover, there was no statistical difference between the benefits Sparky provided and those of Aibo.
“They could pet him, and he would wag his tail and respond to his name,” says Dr. John Morley, a professor of geriatric medicine at St. Louis University and the medical director of one of the nursing homes where the study was done. “Aibo could do what a real dog can, but without pooping, and that’s a major advantage in a nursing home.”
But small studies like these can only be suggestive. Larger clinical trials that randomly assign patients to experimental and control groups are needed before anyone can be sure how much good companion robots can do and how best to use them, Powell says. At least one such trial is underway in Denmark.
But experts say they aren’t particularly surprised that the early results have been promising. And that includes Turkle, who has serious reservations about using robots as companions for the elderly. “It’s not hard to get attachment,” she says. “But I’m saying, attachment at what cost?”
An OK substitute?
While companion robots for the elderly may fill an important social need, they’re not necessarily the best way to do that — or even an acceptable way.
“Are we simply letting ourselves off the hook? That’s always a concern,” says Clifford Nass, a professor of communications at Stanford University and author of “The Man Who Lied to His Laptop: What Machines Teach Us About Human Relationships.” “On the flip side, what do you do for people who have no other option?”
Turkle believes the temptation is great to salve our consciences with robots. “If you leave your mother staring at a wall, you feel terrible,” she says. “If you leave her staring at the TV, you still feel terrible, but a little less. If you leave her talking to a robot that acts like it understands her, maybe you don’t feel great, but better.”
The trouble, of course, is that the robot doesn’t really understand your mother. “Not everything that makes us feel good is actually good,” says Turkle, whose books include “Alone Together: Why We Expect More From Technology and Less From Each Other.”
It’s not that she has a thing against robots. “I love robots,” she insists. “I think they have a tremendous role in elder care.” But she believes that role should be fairly mundane — reaching for things, cleaning things, calling for help.
Robots can only pretend to be companions, she says. We may choose to pretend that they are too. Or not. “Every technology is an opportunity to confront our human purposes.”
Mataric of USC agrees that robots push some ethical boundaries but thinks that they present a reasonable alternative to parking nursing home residents in front of a TV for hours on end.
“At least a robot would make them feel good.” she says. “And if they’re deceived into thinking it cares about them, I don’t think that’s necessarily bad.” After all, people in the prime of their lives have been known to form attachments to their Roombas, “and that’s just a vacuum cleaner.”