The soul of a new machine

Jerome Groop­man writes in The New Yorker about Maja Matarić and the quest to cre­ate robots that can help peo­ple through ther­apy. Assis­tive robots are noth­ing new; Japan’s aging pop­u­la­tion has been using this tech­nol­ogy for years. But Groop­man focuses on the social and emo­tional com­po­nents of these advance­ments, and on efforts to enable robots to under­stand people:

Glanc­ing at the robot, Mary lifted a mag­a­zine from the top of the pile and guided it into a rack on top of the shelf. As soon as the mag­a­zine was in place, the robot emit­ted a beep. Dur­ing the next few min­utes, Mary moved each mag­a­zine, one by one, to the rack. Grad­u­ally, she increased her pace, and the beeps from the robot came faster. Mary began to laugh.

She turned and looked squarely at the robot. With a sly smile, she moved her weak arm toward the remain­ing mag­a­zines on the desk and mimed putting one into the rack. She then stuck her tongue out at the machine.

Matarić said, “She is cheat­ing. She is totally thrilled, because she thinks she cheated the robot.” The robot, though, was on to the game. A reflec­tive white band that Mary wore on her leg allowed the robot to fol­low her move­ments. A thin motion sen­sor attached to her sleeve trans­mit­ted Mary’s ges­tures to the robot, so that it knew almost instantly whether she was rais­ing her arm and in what motion. A sen­sor in the rack sig­nalled the robot when a mag­a­zine was prop­erly placed, and the robot com­mu­ni­cated with Mary only when she per­formed the task correctly.

Although the task lasted about an hour, the nov­elty of the inter­ac­tion did not seem to wane. In a debrief­ing after the study, Mary said, “When I’m at home, my hus­band is use­less. He just says, ‘Do it.’ I much pre­fer the robot to my husband.”

The arti­cle takes an inter­est­ing turn when Groop­man con­sid­ers the ethics of this technology:

Thirty years ago, [MIT pro­fes­sor Sherry] Turkle began study­ing the impact of sophis­ti­cated tech­nolo­gies, includ­ing virtual-reality com­puter games and robots, on emo­tional devel­op­ment in chil­dren and social rela­tion­ships among adults. “I am not a Lud­dite,” Turkle said. “But there is no upside to being social­ized by a robot.” Based on her obser­va­tion of groups of dif­fer­ent ages, Turkle has found that “chil­dren and the elderly start to relate to the object as a per­son. They begin to love it, and nur­ture it, and feel they have to attend to the robot’s inner state.” With this attach­ment and pro­jec­tion of their emo­tions, Turkle says, peo­ple begin to seek reci­procity, want­ing the robot to care for them. “We were wired through evo­lu­tion to feel that when some­thing looks us in the eye, then some­one is at home in it.”

Robots, Turkle argues, risk dis­tort­ing the mean­ing of rela­tion­ships, the bonds of love, and the types of emo­tional accom­mo­da­tion required to form authen­tic human attachments.

In a chat about the arti­cle, Groop­man ties the mat­ter to the expand­ing use of remote-controlled drones in warfare.

Apoc­a­lyp­tic visions involv­ing robots tend to focus on what they’ll do to us. It’s inter­est­ing that the first real anx­i­eties about this rela­tion­ship con­cern what we do with them. Will we become too emo­tion­ally invested? Will we become too dis­tant from the eth­i­cal real­ity of tak­ing human lives?

And I won­der, do Sherry Turkle’s con­cerns extend to things like pets and the Sims?

5 Responses

Tim Carmody says:
November 3, 2009 at 12:30 pm
You know, when some­one uses the phrase “wired through evo­lu­tion,” the chances are pretty good that they’re about to claim some­thing that is at least 150% beyond what you can really justify.
Baby birds are like­wise hard­wired by evo­lu­tion to only accept food from their moth­ers. But these birds can be fooled by pup­pets, which allows zookeep­ers to keep them alive. We might have a thing for eye con­tact from real live human beings, but we can be fooled by movies, pho­tographs, car­toons, pets, and yes, robots. 
There’s almost an alter­nate, weaker ver­sion of the Tur­ing Test at work here: if a per­son inter­acts with some­thing they KNOW to be a robot, and nev­er­the­less talk about “fool­ing them,” stick their tongue out at them, and treat them as an ade­quate social com­pan­ion, then they ARE an ade­quate social com­pan­ion. Grandil­o­quent dis­missals based on a nar­row notion of what evo­lu­tion “ought” to per­mit aren’t just a waste of time; they’re an intel­lec­tual disease.
Reply
Robin Sloan says:
November 3, 2009 at 12:39 pm
Heh heh, I agree. Here’s a fun game:
** ** **
“I am not a Lud­dite,” Turkle said. “But there is no upside to being social­ized by a cat.” Based on her obser­va­tion of groups of dif­fer­ent ages, Turkle has found that “chil­dren and the elderly start to relate to the ani­mal as a per­son. They begin to love it, and nur­ture it, and feel they have to attend to the cat’s inner state.” With this attach­ment and pro­jec­tion of their emo­tions, Turkle says, peo­ple begin to seek reci­procity, want­ing the cat to care for them. “We were wired through evo­lu­tion to feel that when some­thing looks us in the eye, then some­one is at home in it.”
Cats, Turkle argues, risk dis­tort­ing the mean­ing of rela­tion­ships, the bonds of love, and the types of emo­tional accom­mo­da­tion required to form authen­tic human attachments.
** ** **
We have been ascrib­ing nonex­is­tent emo­tional rich­ness to non-human enti­ties for a looong time.
Reply
Matt Penniman says:
November 3, 2009 at 12:51 pm
I agree — and yet…
There is some­thing qual­i­ta­tively dif­fer­ent about a rela­tion­ship with a robot com­pared to a rela­tion­ship with a cat — chiefly, the fact that you can turn a robot off. Robots will never make demands on you when you don’t want them to. They will never require respon­si­bil­ity from you in the way that a cat or (espe­cially) a human would. In that sense, I think Turkle is on to some­thing — a per­son whose major emo­tional rela­tion­ships are with enti­ties that they can utterly con­trol is not a per­son likely to expe­ri­ence much emo­tional growth.
Maybe that’s too nar­row a read­ing of the social pos­si­bil­i­ties of the robot; but it seems like some­thing worth being aware of.
Matt Thompson says:
November 3, 2009 at 1:22 pm
Even while I’m tempted to dis­miss Turkle’s con­cerns, she seems like she’s trou­bled by this after actu­ally explor­ing it; she’s not just a ran­dom pun­dit Groop­man nabbed to com­ment on robots. I give her con­cerns at least as much def­er­ence as I’d give to, say, Nicholas Carr. I’m actu­ally inter­ested in read­ing her book Sim­u­la­tion and Its Dis­con­tents.
I think part of her con­cern comes from the fact that the reha­bil­i­ta­tive robots are actu­ally engi­neered to partly emu­late dis­tinctly and rec­og­niz­ably human char­ac­ter­is­tics, and the arti­cle gets into this. (Groop­man spends a lot of time talk­ing about efforts to avoid the uncanny val­ley.) One of the peo­ple inter­viewed for the arti­cle, for exam­ple, is David Han­son, whose con­tri­bu­tions to the art of robot-making include the inven­tion of a mate­r­ial called Frub­ber, intended to func­tion like human skin. I sus­pect the verisimil­i­tude is part of what makes Sherry Turkle uncom­fort­able, as well as what Matt P men­tions above.
Tim Carmody says:
November 3, 2009 at 2:03 pm
Maybe I mis­read her! But then again, I have always been the snarki­est snarker on the market.