[From Roomba to Kuri to sophisticated and adorable future companion robots that take care of us, we can’t help perceiving machines as independent, sentient social actors – i.e., experiencing media-as-social-actor presence. This interesting story from Wired considers some key design choices and ethical challenges. See the original for a 5:12 minute video. –Matthew]
Companion Robots Are Here. Just Don’t Fall In Love With Them
Matt Simon
August 2, 20.17
“Hey, Kuri,” I say. “I love you.”
Pause. I brace for rejection, but then the robot lets out a balooop and shimmies back and forth. This, I am to presume, means Kuri loves me too.
Interacting with Kuri, a robot set to hit the market in December, is at once fascinating, delightful, and puzzling. Kuri’s creators call it a “companion robot,” but this is no Furby. Kuri belongs to a new class of machines that actually are intelligent, and actually make useful assistants at home. You can see them out in the wild, helping disabled people with routine daily tasks. Soon they’ll remind the elderly to take their medication. Kuri’s more of an all-purpose companion, a member of your family that also happens to play music and take video.
But the vanguard of increasingly intelligent machines invites questions about how people should interact with them. How do we build relationships with what is essentially a new kind of being? How do roboticists make it clear to people that the bond they form with a machine will never be as robust as a bond with a human? And how does the system keep bad actors from exploiting these bonds to, say, use these robot companions to squeeze money out of the elderly?
All big questions that society must start talking about, and now. Sure, no robot is in danger of forming a complex bond with its owner—not even Kuri. The technology just isn’t there yet. But the arrival of Kuri and other companion robots means that in the near future, you’ll need to pay very close attention to how robots make you feel. I mean, I just declared my love for one, for Pete’s sake.
Welcome to the Machines
For a robot without arms, Kuri’s got a lot going for it. You can teach it the layout of your house, then send it to the kitchen or living room to check on the kids. It can navigate bumps with surprising ease. It can play music on command. Using machine vision, it can recognize different members of the family, including pets, and automatically shoot video (a feature known as Kuri Vision). And if you suspect the dog is up on the couch again while you’re at work, you can remote-control Kuri to yell at it.
The really interesting bits about Kuri, though, are its subtleties, particularly when it comes to its interactions with humans. “It’s the little things,” says Mike Beebe, CEO of Mayfield Robotics, Kuri’s maker. “Sometimes like she blinks and she’ll look around, or when she’s about to turn she’ll look first and then she’ll turn. Doing that lets you understand what’s kind of going on inside of Kuri.”
Since Kuri can’t speak—at least, not in what we’d identify as a human language—it communicates with clever cues. A beep means yes and a bloop means no, like a simplified version of whatever the hell R2-D2 speaks. This was an important consideration for Kuri’s designers, because if you want people to get along with companion robots, you have to set expectations for how much they can understand or do. “When something speaks back to you in fluent natural language, you expect at least a child’s level of intelligence,” Beebe says. “That’s a really wonderful thing, if it were possible. But right now robotics just isn’t there yet. So setting that expectation right keeps it more understandable.”
Another companion bot in development, ElliQ, approaches the challenges of human-robot interaction a bit differently. This desktop robot looks vaguely humanoid, with a big noggin that tilts and swivels as it speaks English. It’s meant for the elderly, periodically calling out to remind them to get some exercise or take their meds.
ElliQ works in concert with a tablet to both relay information and express it. If one of your relatives uploads a photo to Facebook, ElliQ will bring it up on the tablet, then turn toward the screen while speaking, in a way gesturing without hands. “We found that the separation between the content on the screen and the entity allows us to do a lot of interactions,” says Dor Skuler, CEO of Intuition Robotics, which developed ElliQ. “We also looked for an aesthetic which (a) we think is beautiful and (b) we think is not intimidating and allows us to earn a right at the home.”
Like Kuri, ElliQ must implicitly set expectations about what it can—and can’t—do. These things don’t have agency and they certainly don’t have consciousness, but the human brain tends to project such things onto robots, especially adorable humanoid ones.
So makers of companion robots want users to have pleasant interactions and make it abundantly clear that robots are tools, not replacements for friends or family. “We’re trying to create an experience that’s more engaging,” Skuler says, “but I think setting the right expectations, that essentially she’s a connector between people. She helps you have meaningful relationships between your loved ones through her, but not with her.”
Our Robots, Ourselves
One day, though, super-sophisticated robots will form complex bonds with humans. Take it from me: I told a robot I loved it. And that’s where the ethics of companion robots get sticky.
Our relationship to robots is wholly different than the relationship we have with, say, our pets. Pets can at least demonstrate nonverbally that they appreciate you by licking your face or bringing you dead animals. (Or, in the case of cats, how eternally indifferent they usually are to you.) But robots? Kuri can’t tell me it loves me back, because Kuri cannot love. It cannot even feel.
These limitations are easy enough for most people to understand now. But as AI gets smarter and smarter, it will be easier to trick people—especially children and the elderly—into thinking the relationship is reciprocal. And such a bond is a powerful thing. Imagine an unscrupulous toy maker inventing a doll so sophisticated that it appears animate to a kid. Now imagine the toy maker exploiting that bond by having the doll tell the kid to buy a personality upgrade for $50.
For the elderly, the target market for some of these new robots, matters will grow all the more complicated. Pretty soon, robots may help them dress or lift them out of bed. “When you’re limited in what you can do yourself and you start relying on the robot, it will lead to a sort of gratitude that then transfers into some sort of attachment relationship,” says Matthias Scheutz, director of the Human-Robot Interaction Laboratory at Tufts. “We see that actually even with simple robots like the Roombas, already, that people become somehow attached to these robots.” DJ Roomba, for instance.
Physically assisting the elderly also introduces an extremely powerful phenomenon: touch. Robots touching humans sounds weird and, well, metallic. Not for much longer, as robots get gentler and literally softer. “They could amplify and further the development of these bonds simply because touch is something very intimate to people, and you might not want that,” Scheutz says. “You might not want the person to build this funny unidirectional relationship with the robot, because the robot cannot reciprocate it.”
Are robots there yet? Absolutely not. But Kuri and ElliQ are your first glimpse at a future where you inevitably form bonds with adorable machines. There’s nothing wrong with that—at least, not yet. But if we don’t manage our expectations, we’re in for heartbreak.