Watch the news, read the tech blogs, and the so-called robots have already taken over.
On Kickstarter there’s a robot that brews beer, a robot that tunes your guitar, a robot that is a 3D printer. There’s a robot refrigerator, a robot alarm clock, a robot desk lamp. In the Internet of Things, bot is the new dot-com, a suffix that confers the mystique of technological innovation on the otherwise mundane. Everything with a few wires in it is apparently robot material.
It’s understandable that inventors, public relations professionals, and even tech journalists would want to call something a robot rather than an alarm clock with wheels. It sounds more complicated, more exotic, more valuable. But isn’t there a difference between a robot and a machine, an automaton, and an appliance?
Unfortunately for the language police, there sort of isn’t. The reason “robot” is such an easy term to abuse is because it’s such a hard one to define. Brad Knox is a robotics researcher at MIT’s Media Lab, and in an e-mail he puts it this way: “The term ‘robot’ … lacks a black-and-white border. Worse, ‘robot’ lacks even a grayscale border that’s consistent across different people’s usages.”
Artificial intelligence researchers tend to use it to refer to machines that are either autonomous or can sense their environment and respond in such a way that affects that environment. Intelligence, of course, is a famously tricky term to define. As Knox points out, a clothes dryer that gauges the wetness of the clothes inside it and adjusts accordingly meets the basic definition. So does a laptop, he adds, “since it ‘senses’ the world through a keyboard, a mouse, etc. and acts through a screen and speakers.”
The blurriness isn’t simply semantic, either. Jan de Ruiter, a psycholinguist at the University of Bielefeld, in Germany, is working on research to create robots that can read facial expressions. Human beings, he points out, are incorrigible anthropomorphizers, and it takes remarkably little to get us to ascribe intention—even personality—to machines. “When a printer doesn’t work, and it’s indicating that the paper is jammed but actually the problem is something else, people talk about how the printer ‘thinks’ there’s a paper jam,” he says.
I looked at some of this research a couple of years ago in an article on videoconferencing robots:
“Pamela J. Hinds, a Stanford University professor of organizational science, has looked at how doctors and nurses interact with HelpMate robots that transport drugs and other supplies around hospitals, and how a team of researchers working in the U.S. treated the remote rover they used to collect samples in the Atacama Desert in Chile. ‘I think one of the things that surprised us the most was the extent to which they were anthropomorphized, even when they were these big, boxy file-cabinet-looking things,’ Hinds says. All a robot had to do was move around in a purposeful way, and people thought of it, in some ways, as a co-worker. People invariably give their robots names and talk about the robots’ moods and tendencies. Hinds recalls seeing one HelpMate to which hospital workers had attached a pair of googly eyes.”
One way to rephrase what Knox says is to say that robots are machines that seem like they’re alive. The reason we let people get away with calling wheeled alarm clocks robots is that, on an instinctive level, we think about the device in those terms.