The SWORDS robot, designed by Foster Mills is still deployed in Iraq, but languishes behind sandbags, dreaming of someday getting some battle action. (Source: Wired: Danger Room)
Turns out the little buggers won't be leaving after all!
After initial reports from Popular Mechanics and other "more reliable" print news sources, that the semi-autonomous SWORDS war-robots were being "yanked" from Iraq due to their accidental targeting of human entities, it turns out that they only got the story half right.
While the robots did indeed get withdrawn from the battlefield, they're still in Iraq being tested and may be redeployed in the near future. This is much sooner than initial quotes by the program manager, Kevin Fahey, indicated (he previously mentioned as span of 10 years or more before the program would be fully active again in case of such an event). Now, Kevin Fahey states more conservatively, "SWORD is still deployed. We continue to learn from it and will continue to expand the use of armed robots."
Foster Miller, maker of the SWORD robot, was quick to try to dispel the hype that surrounded the announcement that the robots might be shooting at our soldiers, originally propagated by Popular Mechanics. Said spokeswoman Cynthia Black, "The whole thing is an urban legend."
Black helpfully provided information on the specific cases in which "uncommanded movement" occurred. She stated, "One case involved a loose wire. So, now there is now redundant wiring on every circuit. One involved a solder, a connection that broke. Everything now is double-soldered."
The third case had nothing to do with a gun. The robot was placed on a 45 degree incline and left to run for two and a half hours as part of an endurance test. Predictably, the motor burned out and overheated. She states, "When the motor started to overheat, the robot shut the motor off, that caused the robot to slide back down the incline. Those are the three uncommanded movements."
There does seem to be some confusion between Fahey and Foster Mills, as Fahey was reported as saying the robots recently did something "very bad" (indicating a serious recent failure).
While Black and Fahey fail to mention it, one thing that makes a "killer robot" scenario unlikely is the fact that the robots really are about as green as it comes when it comes to combat. They (by all reports) are not being used very much. Furthermore, they have not notched even one kill of an enemy combatant.
Stew Magnuson, a reporter for National Defense, was at the same news conference as Popular Mechanics, and helped to offer a bit more clarity as to what is really going on. The real scenario he offered is that the robots are languishing in under-deployment to the frustration of Foster Mills and the conference was largely an effort by Foster Mills to express their disappointment in this situation.
Robert Quinn, vice president of Talon operations at Foster-Miller, stated at the news conference that the three robots sent to Iraq were stuck in stationary positions behind sandbags and not out patrolling the streets as its designers planned. Army leadership was fearful the "through an interface" targeting of the robots via human operators in combat situations would produce dangerous results, and thus refused to deploy them.
Quinn complained, "If you have a mobile weapons platform that can’t be mobile, and it becomes nothing more than a fixed position, then why not just put it on a tripod."
The Army argues, he said, that they have not addressed and trained in the tactics, techniques, and procedures for using the robots in battle. However, they say they can't put such measures in place until there are a sufficient number of SWORDS robots in place. Foster Miller says they can deploy more robots, necessarily, without the information obtained from the trial deployment, so it’s a vicious case of the "chicken or the egg" for the Army and Foster Mills.
The SWORDS robots are controlled via a human operator on a terminal, but are semi-autonomous in that they can perform basic target acquisition. They carry M249 light machine guns. While the robots have been somewhat jokingly portrayed by the media as possible "rogue killers", in reality the greatest chance for failure rests in the hands of the human operator. Likely, the largest risk of friendly fire stems from if the cameras on the robot provide confusing or insufficient information during a combat situation and the human operator errantly fired as a result.
While the SWORDS are relatively germane, the concept of fully autonomous war robots, using the advanced targeting algorithms that are currently being researched, is still a fair subject of intense debate. It will likely be tempting in terms of resources saved for the military to eventually take the human operator out of the equation, given sufficient advances in artificial intelligence and mechanics.
This decision is not a matter of if, as much as when. When it occurs, the closing point of the prior article stands -- despite all the testing and development effort possible, is it truly possible to entirely rule out the chance of the robot turning on its human controllers?
