Progress in the domain that Marvin Minsky once characterized as "making machines do things that would be considered intelligent if done by people" has not been as dramatic as its founders might once have hoped, but the penetration of machine cognition into everyday life (from the computer that plays chess to the computer that determines if your toast is done) has been broad and deep. We now the term "intelligent" to refer to the kind of helpful smartness embedded in such objects. So the language has shifted and the question has disappeared. But until recently, there was a tendency to limit appreciation of machine mental prowess to the realm of the cognitive. In other words, acceptance of artificial intelligence came with a certain "romantic reaction." People were willing to accept that simulated thinking might well be deemed thinking, but simulated feeling was not feeling. Simulated love could never be love.
These days, however, the realm of machine emotion has become a contested terrain. There is research in "affective computing" and in robotics which produces virtual pets and digital dolls — objects that present themselves as experiencing subjects. In artificial intelligence's "essentialist" past, researchers tried to argue that the machines they had built were "really" intelligent. In the current business of building machines that self-present as "creatures," the work of inferring emotion is left in large part to the user. The new artificial creatures are designed to push our evolutionary buttons to respond to their speech, their gestures, and their demands for nurturance by experiencing them as sentient, even emotional. And people are indeed inclined to respond to creatures they teach and nurture by caring about them, often in spite of themselves. People tell themselves that the robot dog is a program embodied in plastic, but they become fond of it all the same. They want to care for it and they want it to care for them.
In cultural terms, old questions about machine intelligence has given way to a question not about the machines but about us: What kind of relationships is it appropriate to have with a machine? It is significant that this question has become relevant in a day-to-day sense during a period of unprecedented human redefinition through genomics and psychopharmacology, fields that along with robotics, encourage us to ask not only whether machines will be able to think like people, but whether people have always thought like machines
SHERRY TURKLE is a professor of the sociology of science at MIT. She is the author of Life on the Screen: Identity in the Age of the Internet; The Second Self: Computers and the Human Spirit; and Psychoanalytic Politics: Jacques Lacan and Freud's French Revolution.