Exterminate! Exterminate!

When a Thai engineering professor announced that his students had created the "world's first armed robot guard," controllable via the Internet, the global scientific community shuddered in unison. Its inventors, however, swear it's a useful security device. Would the robot ever be sold commercially?

Kevin McLaughlin

Robots have always captivated the imagination--just consider the number of movies and video games in which they wreak havoc. But last month, when a Thai engineering professor unveiled an armed security robot created by students, scientists, and robotics experts around the world were not charmed. 

The robot--dubbed "Roboguard" by the Thai press--is designed to take the place of human security guards in restricted areas such as banks and top secret military zones. Equipped with a video camera, Roboguard can track the movements of intruders using a built-in infrared sensor. But locating intruders isn't the trick up this robot's metallic sleeve. Roboguard, you see, is packing heat. The robot comes equipped with a gun, which can be manually aimed and fired via the Internet or Local Area Network (LAN). Understandably, this feature is attracting the most attention. 

Roboguard's inventor, Dr. Pitikhate Sooraksa of King Mongkut's Institute of Technology in Bangkok, says his invention is incapable of violating robotics-visionary Isaac Asimov's three laws of robotics. One of the laws states that robots must never injure human beings. "Some people are saying that the Roboguard can think for itself and shoot whenever it wants, but that is simply not the case," Pitikhate says. "We have taken care to build the robot so it cannot fire on its own." 

Dr. Pitikhate can't understand the fuss, and points out that the U.S. military has been developing sentry robots for many years, some of which have lethal capabilities. "We've used tele-robotic technology that already exists," he says. "All we did was add a pistol and a camera, as well as the ability to connect it to a network." 

Military sentry robots primarily watch over areas like warehouses and sound an alarm when they detect something wrong, like an intruder or a fire. Most military "robots" used in combat consist of computer chips with navigation instructions, for instance, in smart missiles. These robots always have a human in the loop who makes the decision to fire. "It has long been the policy of the U.S. military not to trust robots with weapons," says Chuck Thorpe, acting director of the Robotics Institute at Carnegie Mellon University. 

It's Roboguard's link to the Internet or a Local Area Network (LAN) that raises scientists' suspicions. For instance, what happens if network congestion occurs at the precise moment that Roboguard is locking onto its target? "Although the command to fire is made up of very few bytes [which means that it travels across the network quickly], the tests we did using the Internet were purely experimental," says Dr. Pitikhate. "If Roboguard were to be used in an area such as a bank vault or a military installation, it would require a LAN [to overcome the time delay]." 

Dr. Pitikhate never expected that the project would come under such intense scrutiny. "The goal of the Roboguard project was to challenge my students and help them become skilled engineers," he says. "This was a scientific experiment that we never intended to market commercially." 

He feels that some of the negative media attention is unwarranted. "These types of robots are a necessary evil for the defense industry," he opines. "If people are trying to get into your top secret places in order to destroy whatever's in there, should you just open the door and let them in?" 

Kevin McLaughlin(kmclaughlin@business2.com)is a reporter for Business2.com