In April, the government of Japan released more than 60 pages of recommendations to “secure the safe performance of next-generation robots,” which called for a centralized database to log all robot-inflicted human injuries. That same month, the European Robotics Research Network (EURON) updated its “Roboethics Roadmap,” a document broadly listing the ethical implications of projected developments like robotic surgeons, soldiers, and sex workers. And in March, South Korea provided a sneak peek at its “Robot Ethics Charter” slated for release later in 2007. The charter envisioned a near future wherein humans may run the risk of becoming emotionally dependent on or addicted to their robots.
The close timing of these three developments reflects a sudden upswing in international awareness that the pace of progress in robotics is rapidly propelling these fields into uncharted ethical realms. Gianmarco Veruggio, the Genoa University roboticist who organized the first international roboethics conference in 2004, says, “We are close to a robotics invasion.” Across the technologically developed world, we’re building progressively more human-like machines, in part as a result of a need for functional, realistic prosthetics, but also because we just seem to be attracted to the idea of making them. Honda’s ASIMO and Sony’s QRIO are the remarkable proof-of-concept products that illustrate a strange yet pervasive urge in us to build ourselves all over again.
It is the social interaction of these and similar machines that raises the most interesting questions, however. “Social robots” are now entering human culture, most frequently as entertainers for the very young and as caretakers for the very old. In Japan, consumers buy “therapeutic” robots like the humanoid Wakamaru, which is designed to provide companionship for the elderly and disabled and is capable of rudimentary social interactions with its owners. In the US, recent holiday seasons have seen parents clamoring for Furbys, Tickle Me Elmos, and other robotic toys for their kids to “nurture” and play with. It is this drive to build robots that appear to understand us and engage with us—and perhaps one day think like us—that is providing scientists with some unsettling and unique insights. And it’s driving the emerging field of roboethics, which asks questions about how these machines affect us and how best to integrate them into our culture.
Of course, we’ve been grappling with the idea of physical and emotional dependence on our artificial creations since at least the time of the Romans. In the poet Ovid’s story of the lonely sculptor Pygmalion, the artist becomes infatuated with his lifelike creation, the feminine statue Galatea. In Ovid’s original tale, the gods bring Galatea to life and the couple conceives a son, but later versions take darker twists: George Bernard Shaw’s 1913 play, Pygmalion, portrays an emotionally distant Eliza Doolittle (Galatea) ultimately rejecting her creator’s affections.
A scientific understanding of human response to social robots began with MIT computer scientist Joseph Weizenbaum’s landmark experiments in 1966. Weizenbaum had developed a computer program that crudely mimicked a psychotherapist by rephrasing statements from human “patients” back to them as questions, thus supportively reflecting their thoughts. A user input of “I feel frustrated,” for instance, returned, “Why do you feel frustrated?” Weizenbaum named his program ELIZA after the Galatea character in Shaw’s play, whose mimicry of aristocratic speech propels her into high society.
Weizenbaum was deeply troubled by what he discovered during his experiments with ELIZA: Some of his students exhibited strong emotional connections to the program; some actually wished to be alone with it. Weizenbaum had unexpectedly discovered that, even if fully aware that they are talking to a simple computer program, people will nonetheless treat it as if it were a real, thinking being that cared about their problems—a phenomenon now known as the “Eliza Effect.”
Brian Scassellati, a roboticist at Yale University who is one of the leading researchers looking at what robots can teach us about human social interaction, points out that it takes very little for even the most basic social robots to elicit a response from us that is, at its core, deeply human in nature. “The highest-level effect we see over and over again is that people really want to treat these things as if they were human… we want to treat them as though they have a lot more knowledge and capability than they have.”
When he was a grad student, Scassellati was working on gaze behavior, programming a robot to look at anything that moved. During testing, the machine would see his hands move at his computer. “It would look down at me,” he says, “and of course, then I’d look up at it. Eventually, it would habituate to me, but then 30 seconds later, it would look back at me again, and I’d look back at it. This would go on for hours at a time.” Scassellati had to work for months to train himself not to instinctually return the gaze. “There was this very basic social behavior in me saying, ‘Someone just turned to look at you; you’d better see who it is.’ When we see that generated in us as designers, we know we’ve done something very right; we’ve hit something that’s very basic and primitive.”
Social scientist Sherry Turkle, the director of MIT’s Initiative on Technology and Self and one of Weizenbaum’s former colleagues, calls ELIZA and its ilk “relational artifacts”: machines that use simple tricks like mirroring speech or holding eye contact to appeal to our emotions and trigger a sense of social engagement. In 2007, more than 40 years after ELIZA, the designers of machines like ASIMO and Wakamaru have only further enhanced the potential for the Eliza Effect to engage us. “The relational artifacts of the past decade, specifically designed to make people feel understood, are more sophisticated interfaces, but they are still parlor tricks,” Turkle says.
And so, roboethics is starting to ask some questions for which we, as yet, have no concrete answers. “If our experience with [these robots] is based on a fundamentally deceitful interchange—[their] ability to persuade us that they know of and care about our existence—can it be good for us?” asks Turkle.
Scassellati points out that the effects of social robots move beyond the psychological; there is a sociological effect on us as a culture. “There was a huge outcry when Sony decided not to continue producing the AIBO [a doglike, social robot], in part because there were many people who were very attached to the device—almost as attached as they’d be to a pet,” he says. “So what happens when these devices become more a part of our lives? We just don’t know the answers at this point.”
The three recent roboethics charters—EURON’s and those of Japan and South Korea—broadly aim to ensure that as we incorporate these emotionally powerful machines into society, we do so in ways that minimize any harm their introduction might cause.
The current increase in our cultural awareness of, and research into, robot-human interaction is only an early step in our evolving relationship with these social machines of our own making. We’ve created incredibly advanced tools that help us by fooling us into feeling comfortable with them: We’ve come a long way since Pygmalion, and yet, this is only the beginning.
Originally published July 16, 2007