One doesn’t have to be a fan of Philip K. Dick to recognize that robots are already participating in our daily routines to the extent that their activities have legal implications. That will become more important as they act with more autonomy, and the law will have to change to deal with it.
The latest edition of Science features a multi-part report on the current state of robotics and its future. Leaving aside the discussions therein of pure technology, the most fascinating article is an interview with Ryan Calo of the University of Washington law school. Calo’s forthcoming paper on “Robots and the Lessons of Cyberlaw” will appear in a forthcoming issue of the California Law Review.
Calo observes that robots have a way of undermining the law’s clear distinction “between a thing and a person.... You get compensated differently when someone else’s negligence results in injury to property than to a person. When property is involved, you get market value. With people you have been deprived of that person’s companionship. To the extent that people become heavily enmeshed socially with robots, the law will have to decide into what category to sort them.” Calo doesn’t foresee the creation of “stand-alone rights” for robots, even to the extent the law has begun to recognize animal rights. Rather, he sees the law extending the rights and responsibilities of owners to their robots -- if one uses a software agent or “bot” to make a deal, for example, the owner should be held to its terms unless something about it looks “objectively implausible.”
As robots become more lifelike and integrated into our social interactions, Calo says, the law may begin to treat human-robot interactions as indicators of human predispositions. Today, he observes, some police officers are instructed that if they respond to an animal abuse complaint in a home where children are present, they should alert child welfare services. “You could imagine tweaking those policies to apply if you were to have reports that someone was kicking their robot dog.”
Calo is not the first expert to ponder the future of robot-related law. Last year, Neil M. Richards of Washington University in St. Louis and William Smart of Oregon State examined how the law would have to treat liability issues as robots became more autonomous and entered “public life and our private homes.”
The most adept robots today, such as the Mars Rover, tend to operate with shared autonomy, Richards and Smart observe. “They receive high-level instructions from human operators (‘go over to that boulder’), but are responsible for their own low-level behavior (avoiding obstacles, for instance).
That opens a gray area of liability: if a semiautonomous agent screws up, they ask, who’s responsible? “Is it the manufacturer, the programmer, the user (who gave a bad instruction) or some combination of them all?”
With the advent of self-driving cars -- and Elon Musk’s Tesla Motors already incorporates in some of its models an autopilot technology for everyday driving -- the question is no longer academic or fanciful. If an auto-driven Tesla has an accident, who do you sue? (Musk says the technology is autopilot but not “autonomous”: “It’s not at the level of safety where you could just fall asleep and arrive at your destination.”)
Richards and Smart do warn against making too much of robotic autonomy; to a large extent, it’s a myth fostered by the “android fallacy”: because a robot may resemble a human and respond in subtly different ways to different inputs, they may seem to have free will, when they’re really just executing software instructions.
“Robots are, and for many years will remain, tools,” they caution -- “no different in essence than a hammer, a power drill, a word processor.” Legislators may be fooled into holding designers of human-like robots less responsible than those of “robotic” robots. The point, they say, is to hold designers and manufacturers to the same high standards that apply in the automobile industry. Cars can be anthropomorphized and appear to have lives of their own, but they’re still only dumb mechanisms. It will be a long time before robots, no matter how we may think they “feel” or “see,” are anything more than that.