RSS
Is teaching robots to lie a bad idea?
Scientists have taught robots how to deceive others, prompting fears of a Terminator 2–style apocalypse
A robot that can lie would excel in the political arena, some say.
A robot that can lie would excel in the political arena, some say.
Corbis
I

n a scientific breakthrough that critics are portentously calling a "terrible, terrible mistake," researchers at the Georgia Institute of Technology have taught robots how to lie. Here's a guide to the simple experiment that some worry has hastened the rise of the machines:

How did this experiment work?
Ronald Arkin and Alan Wagner, researchers at Georgia Tech, taught a robot to make it look as if it had hidden in one location when it was actually hiding in another. The robot created a "false trail" by knocking over markers for another robot to find, and successfully deceived the second robot 75 percent of the time. The resulting study appeared in the International Journal of Social Robotics.

What use could we have for a lying robot?
It might have applications for military technology, say the study's authors. "A search and rescue robot may need to deceive in order to calm or receive cooperation from a panicking victim," they say. They also point out that camouflage is a form of deception.

Isn't this dangerous? Have they never seen Terminator 2? 
Apparently not, says Duncan Geere at Wired. Sure, robots able to trick battlefield foes may seem like useful technology in principle. "But when machines rise up against humans and the robot apocalypse arrives," Geere says, "we’re all going to be wishing that Ronald Arkin and Alan Wagner had kept their ideas to themselves." 

What do the researchers say about the apocalyptic implications of their research?
The pair admit that their methods could be used for "nefarious purposes," and urge robotics communities to come up with ethical guidelines on robot deception for the "well-being of society." It's worth noting, says Andrew Moseman at Discover, that Arkin is also "one of the chief minds mulling over the moral quandaries of granting machines more autonomy." The robotics expert has even suggested the design of an "ethical governor" for fighting future wars — a set of programmed rules to prevent robots from massacring civilians or desecrating religious buildings.  

Are there any other practical uses for a lying robot?
They could run for political office, says Tommy Maple in the Independent Florida Alligator. After all, "hypocritical androids" would do a much better job in the modern-day political environment than "weak humans with their annoying gray areas and tendency toward consensus." 

Sources: The International Journal of Science Robotics, Wired, Discover, Smart Planet

EDITORS' PICKS

THE WEEK'S AUDIOPHILE PODCASTS: LISTEN SMARTER

Subscribe to the Week