One major goal of contemporary robotics is to create sociable robots that have emotions and facial expression, to make it easier for humans to relate to them. If robots and humans can interact with one another more effectively then robots will be used more in homes and offices. The science fiction cliché of the rational robot without emotions is what AI programmers are trying to avoid. The ideal might be something more like Data, a robot that can learn to approximate human emotions.
The first steps in creating sociable robots involve developing robots that face you when you speak and have realistic facial expressions. Ray Kurzweil’s Singularity site has a video and links to “Affetto, a Child Robot with Realistic Facial Expression” (http://www.kurzweilai.net/meet-affetto-a-child-robot-with-realistic-faci...). I actually find Affetto to be a bit scary, especially when they remove his face in the videos.
Interactions with computers could be heightened if computers had sensors which could tell an AI when you’re getting bored or frustrated. If computers pay attention to our mental or physical states, then a video game could become more challenging when we get bored. Or a computer could try to alter its actions because we’re annoyed. Sensors already exist that refuse to start a vehicle because the driver is intoxicated.
The Chronicle of Higher Education has an article, “Programmed for Love,” about Sherry Turkle, a psychologist who has spent several years studying sociable robots (http://chronicle.com/article/Programmed-for-Love-The/125922/). Once an advocate for sociable machines, Turkle has found herself increasingly skeptical about the effect of sociable machines on humans.
In Turkle’s new book, Alone Together: Why We Expect More From Technology and Less From Each Other, she claims that we are losing our ability to be ourselves and relate to other people as we become more and more dependent on our technology and let our cellphones and laptops dominate more of our lives. Turkle points out that it’s easy for people to become emotionally attached to robot dogs and electronic dolls even if those objects cannot actually love us back.
The article also refers to David Levy’s recent book Love and Sex with Robots, which predicts that in the near future that people will be having sex with robots and falling in love with them, even marrying robots by 2050. Levy suggests that you would be better off having a robot babysit your child rather than an inexperienced teenager.
Turkle is opposed to robots taking on caretaker jobs like nursing home attendant, babysitting, or as companions for people with disabilities because those jobs should be held by humans, who can genuinely reciprocate emotions, not just approximate them.
The article profiles an intriguing dispute between Levy and Turkle. Levy believes that some adults will be happier in a relationship with a robot because they lack the social skills to develop a relationship with a person. Lonely people don’t have to be miserable; they can have a robot companion. Turkle believes that a shy, awkward person would be better off if they developed the social skills to have a relationship with a real person.
The task of creating sociable robots is perhaps more advanced in Japan than anywhere else. Japan has long had the reputation of being the land of the robot, where people have a much more positive image of humanoid machines than the West, where robots are often portrayed as Terminators.
The Japanese have been trying to develop sociable robots that will address a growing concern for them: care for the elderly. In Japan a falling birthrate and increasingly long lives means the country has an aging population with a pension crisis and a worker shortage. Japan is reluctant to admit more immigrants, feeling that immigration would destroy their culture, so the country is trying to fill the labor shortage with robots.
But the Japanese have run into a problem. According to Popular Science (http://www.popsci.com/technology/article/2011-02/some-japanese-patients-...), the elderly don’t like being cared for by robots. The robots can’t do very much yet, and the elderly prefer to interact with humans.
The article is brief and doesn’t go into much detail, but robots would be cheaper than human caretakers. Also, as the Japanese population ages, and they experience a shortage of human caretakers, those workers will become more expensive; it’s a basic economic law that scarcity drives up prices. The Japanese will be forced to either develop effective robots or allow more immigration.
I’m curious how our readers feel about this. Would you like a robot caregiver if you were sick or old? Is Turkle right that a relationship with a robot caregiver is more impoverished than a relationship with a human? The immediate response to the idea of a robot caregiver is that it’s cold and inhuman. But have you ever been in a nursing home and seen how the staff treats the elderly? I’d rather be taken care of by a robot programmed to care about my needs than a sadistic or nasty attendant. How about a robot babysitter? Would a robot be better than a teenager who sneaks her boyfriend in while you’re out?
Sherry Turkle has conducted a great deal of research – hundreds of interviews – that have shown her that humans can develop a deep emotional connection with even the primitive robots that now exist, such as robot dogs and dolls. A number of people in studies have formed attachments to robot dogs because they’re cute and bark. But unlike a real dog, the robot doesn’t reciprocate your affection. Even if you program a robot to appear to care, the emotions aren’t genuine. Those points are largely true, but if I were elderly I’d rather be taken care of by a robot that pretends to care, rather than a annoyed, distracted human ready to get off work and deal with his/her angry spouse or demanding children.
Also, the question of “pretending to care” is an interesting one that will, once robots are more advanced, become quite complex. If a complicated future AI – one that could pass the Turing test -- is programmed to have emotions, isn’t it possible that it then has emotions? Robots will think and feel on silicon, which is clearly different than thinking on a biological brain, but that doesn’t mean the thinking and the emotions aren’t real. After all, humans have evolved to behave in certain ways, which is a type of programming, perhaps just as limited as the programming robots will receive.
Part of the reason the current generation of the elderly reject the robots is because the current generation of robots aren’t that advanced and useful. In another 10-20 years robots may be much more effective caregivers. Furthermore, future generations of the elderly may be more comfortable with technology than the elderly are now and more willing to accept robonurse.
Robert Bee is a professional librarian and a freelance writer living in New Jersey. He can be reached at email@example.com