After Fox News misrepresents a military robot’s dietary habits, the world muses over what ethical behavior means for intelligent programs and machines.
Ever since news broke earlier this month about the Energetically Autonomous Tactical Robot, or EATR, news outlets have been racing each other to weigh in on the coming robot apocalypse. This week, people all over the media landscape are suddenly realizing that ostensibly intelligent machines are among us—building our cars, vacuuming our carpets, even entertaining our pets and serving us delicious adult beverages—and that it won’t be long before they realize they have nothing to lose but their chains (or power cords).
Of course, the story that started this media sensation was a tempest in a teapot. EATR, a DARPA-funded joint venture of Robotic Technologies and Cyclone Power Technologies, was designed to “find, ingest, and extract energy from biomass in the environment” in order to work long missions without having to be manually refueled. But news outlets, notably Fox News, interpreted this factoid to mean that it would consume the corpses of fallen soldiers and—who knows?— maybe even the occasional live one.
This prompted the robot’s designers to issue what may be the greatest press release of all time, complete with references to the Geneva Conventions’ prohibitions against the desecration of the dead. The release prompted a predicable response from PETA; the group complimented the robot on its choice of the vegetarian lifestyle, further cementing the impression that the organization is, in fact, a PSYOPS unit of the meat lobby.
Capitalizing on the buzz over malevolent machines, on Wednesday, the Huffington Post ran a months-old robot-bites-man story. The piece described a 2007 incident in Sweden in which an industrial robot mistook a worker’s head for a heavy rock, leading to a $3,000 fine against the robot’s manufacturer.
Some serious discussion and debate did emerge from this heightened awareness of robotics: a New York Times article by John Markoff provided some insight into a February closed-door meeting held by the Association for the Advancement of Artificial Intelligence (AAAI) on the ethical and legal implications of AI, robotics, and the potential for the singularity, that nerdish nirvana where super-intelligent machines become the dominant decision-making force on the planet. Acknowledging that the pubic understanding around these topics largely consists of science-fictional accounts of robot uprisings, the attendees discussed the safeguards that may be necessary to ensure future AIs can be controlled by humans. A report on their conclusions is due out later this year.
But before this meeting is interpreted as a harbinger of unstoppable killing machines, it should be noted that the current problem with robots and the “intelligence” that we give them is not that they are too smart, but too stupid. The Swedish industrial robot is actually a good example of this. A $3,000 fine for a nearly crushed head may seem like a pittance, but in no way was that accident the robot’s fault. The lifter has a well-defined set of behaviors but nothing analogous to human judgment; failing to ensure the robot was totally deactivated was as much a human error as servicing a lawnmower without first disconnecting its power supply.
Related articles by Zemanta
- What happens when artificial intelligence gets too advanced for its own good? (crunchgear.com)
- Want responsible robotics? Start with responsible humans (scienceblog.com)
- The Missing Soldier (andrewsullivan.theatlantic.com)
- ‘Flesh-eating robot’ is vegetarian (guardian.co.uk)