Robot warriors have already seen action in Iraq, and the US Army plans to replace one-third of its armored vehicles and weapons with robots by 2015. These killing machines may one day come equipped with an artificial conscience — even to the extent of disobeying immoral orders.
The US Army’s latest recruits are 1 meter (about 3 feet) tall, wear desert camouflage and are armed with black M249 machine guns. They also move on caterpillar tracks and — thanks to five camera eyes — can even see in the dark.
The fearless fighters are three robot soldiers who, unnoticed by the general public, were deployed in Iraq in mid-June, charged with hunting down insurgents. As if guided by an unseen hand, they hone in on their targets and fire at them with their machine guns. It’s the future of war — and it’s already here.
“It’s the first weaponized robot in the history of warfare,” says Charles Dean, an engineer with Waltham, Massachusetts-based Foster-Miller, the manufacturer of the new devices. Dean and the 70 employees in his department are eager to find out how their three protégés are holding up on the front. Because the three robots, dubbed “Swords,” are being used in a secret mission, their creators have no idea whether the devices have already killed enemy fighters in combat.
FROM THE MAGAZINE
Find out how you can reprint this DER SPIEGEL article in your publication.
It seems only a matter of time before the three combat robots will get some reinforcements. The American military is currently testing the Gladiator, an unmanned mobile device developed by engineers at the Carnegie Mellon University Robotics Institute in Pittsburgh, Pennsylvania. Gladiator weighs more than a ton and comes equipped with rubber tires that enable it to scurry up inclines of up to 60 percent. The US military has already mounted a targeting camera and a remote-controlled M240 machine gun on a prototype.
“We’ve already done plenty of shooting with the machine gun,” says Col. Terry Griffin, who heads the joint US Army and Marine Corps robot program. If further tests are successful, a four-wheel version of the Gladiator could be headed for Iraq next year — assuming US troops are still in the country.
According to Griffin, the combat robot is capable of disbanding groups of undesirables. There are three stages of escalation: First Gladiator issues warnings through a loudspeaker, then it fires rubber bullets and, finally, it starts firing its machine gun.
A New Kind of War
More than 50 years after author Isaac Asimov argued in his classic novel “I, Robot” that a robot should never be allowed to do harm to people, the development of automated killers has become unstoppable. Swords and Gladiator are the harbingers of a new type of warfare, in which killing will increasingly be left up to machines.
The US Army is developing a number of warrior robots.
The US Army is developing a number of warrior robots.
According to an internal US Army memo, armed machines “are making their way onto today’s battlefields and will be extremely widespread on the battlefields of the future.” The Pentagon’s budget already includes up to $200 billion for a modernization program dubbed “Future Combat System.” Under the program, robots will replace one third of armored vehicles and weapons by 2015.
Automated warfare is also making inroads in Israel, where the military deploys robots along the country’s 60-kilometer (37-mile) border with the Gaza Strip. The stationary “See-Shoot” system developed by Rafael, an Israeli weapons manufacturer, includes machine guns and cameras, and has a range of 1,500 meters (4,921 feet).
From a military standpoint, there are many reasons to support the growing use of steel soldiers. For one, fear and fatigue are non-issues. Robots kill without hesitating and, unlike flesh-and-blood soldiers, losing them is merely a financial loss. A new Swords goes for about $150,000. Besides, politicians and generals no longer need to worry about a public outcry over excessive fatalities: Who mourns a fallen tin soldier?
Still in Control
However, human operators are still strictly in control of these mechanical soldiers. It will be a while before the humanoid murderers portrayed in Hollywood films like “Robocop” and “Terminator” will be unleashed on humanity. “But there are no scientific barriers standing in the way of autonomous combat robots,” says Ronald Arkin of the Atlanta-based Georgia Institute of Technology. “The parts of the whole are being assembled as we speak.”
In his corner office, graying robotics expert Arkin investigates ways to prevent the grim scenarios of science fiction films from becoming reality. He is currently conducting an Internet survey in an attempt to determine how military officials, politicians, robotics researchers and ordinary citizens feel about autonomous killing machines.
What are the ethical rules these machines should follow when they are sent into war, for example? To help tackle the issue, Arkin is developing software that could be used to program the machines with such rules — a sort of conscience for steel soldiers.
One of those machines, BigDog, provides a sense of how far robotics has come technologically. The headless device feels its way as it moves forward. A built-in computer and internal sensors ensure that BigDog remains firmly on all fours, even when given a firm kick in the side. The robot, developed by Boston Dynamics, will likely begin its military career as a packhorse.
At a show last week at Webster Field, a military base in Maryland, a craft with a diameter of only 33 centimeters (13 inches) could be seen flying through the air and landing on spring-like legs. US troops in Iraq are currently testing about 20 of these so-called Micro Air Vehicles, which are made by Honeywell Aerospace. The soldiers can either control the drone with a joystick or program it to run on automatic pilot. To do so, they call up a digital map of their surroundings on a computer screen and click on the target. The drone then uses GPS to locate the target.
An even more impressive device on display at Webster Field was a seven-meter (23-foot) helicopter called Fire Scout. Instead of a cockpit, the unmanned helicopter has a windowless face that covers a Cyclops-like eye: a laser device that enables Fire Scout to land on its own, even on the tight deck space available on smaller warships. Fire Scout, of which US manufacturer Northrop Grumman has only produced two prototypes, is still unarmed. But that too will change, says engineer Doug Fronius: “There are definite plans to integrate weapons into the system.”
Northrop Grumman is also developing an unarmed stealth fighter, the X-47, which the company expects to perform its first fully automated landing on a moving aircraft carrier in 2011. “By removing the pilots, we enable the device to remain airborne for an additional 10 hours or more,” says Tighe Parmenter of Northrop Grumman. “To program an enemy mission, all you need is a keyboard and a mouse.” In early August the US Navy awarded the company a contract worth $635.8 million to develop the fighter drones.
In general, airborne robotic devices are the vanguard among military robots. Unmanned flying objects have been used in war zones for some time, mainly for reconnaissance, but also to deliver deadly weapons. The two missiles that killed Al-Qaida terrorist Mohammed Atef in November 2001 as he was traveling to Kabul by car were fired from a Predator drone.
Regardless of whether robots hurtle through the air or serve as mechanical infantrymen on the ground, until now human operators have decided whether they are permitted to shoot. The fear that the machines could suddenly start letting loose on their own troops is still too great.
Before Swords fires its first salvo at terrorists in Iraq, it needs the permission of two human operators. A supervisor presses a button on his remote control, which makes the machine gun operational. At the same time, another soldier must activate two red switches on this control unit to allow the robot to begin shooting.
However, it is only logical that decisions over life and death will increasingly be transferred to the machine — just as soon as engineers have figured out how to overcome the problem of distinguishing between friends and foes. The first device likely to be capable of making this distinction could be installed by as early as this year along the 248-kilometer (154-mile) demilitarized zone between North and South Korea.
Samsung Techwin, a South Korean electronics firm, heads the consortium that developed the device, a black shooting robot. Equipped with video and ultrasound cameras, the robot can distinguish between trees and people and, according to Arkin, can independently open fire on anyone crossing the border illegally.
The Pentagon also wants to give the robots more freedom, arguing that the only way to enhance the fighting power of US troops is to enable a soldier to use several unmanned systems at the same time. This is only possible if the machines are allowed to make many of their decisions independently. According to a US Army document, both “lethal and non-lethal combat” could be possible as autonomous behavior.
The Future is Bright
The attack of the killer robots may sound like some macabre vision of the future. But robotics visionary Arkin also believes that there could be some positive aspects to the scenario. Contrary to many international treaties and declarations of intent, atrocities and human rights violations have always been part of wars in the past. A case in point is the torture scandal involving US troops at Iraq’s Abu Ghraib prison.
Arkin believes that combat robots would not be tempted to commit such atrocities, thanks to their artificial conscience. “Robots could behave more humanely than human beings,” he says. He plans to present a prototype of his morality software at a conference in September. Depending on the situation and its mission, a robot would select, from a wide range of options, the one action that it considers especially ethical — even if it means refusing to obey a command.
Arkin also hopes that the mere presence of unmanned systems could make crises and conflicts more humane. Wherever their cameras are pointed, the robots create a record that could ultimately be open to public scrutiny. Soldiers can then expect everything they do to be captured on camera, an effective deterrent against those who seek to exact revenge and indulge their torture fantasies on prisoners.
In the meantime, the robots are being welcomed with open arms by their human fellow soldiers. The US Army will decide in October whether to deploy additional Swords robots in Iraq. If the soldiers had their way, they would get another 20 Swords, says Michael Zecca of the Picatinny Army Arsenal in New Jersey.
Sign up for Spiegel Online’s daily newsletter and get the best of Der Spiegel’s and Spiegel Online’s international coverage in your In- Box everyday.
Soldiers are especially fond of the hundreds of unmanned robots that have been used for years as minesweepers in Iraq and Afghanistan, and are credited with saving countless lives. The small tin soldiers are so valuable to the military that they are even promoted and decorated with medals. Whenever a mine detonates under one of the devices, soldiers prefer to repair the robot if at all possible, rather than have it replaced with a brand-new substitute.
At a military base in Yuma, Arizona, a colonel ordered soldiers to break off a test in which a robot was being repeatedly sent into a minefield — because, so he said, it was inhumane.
Arkin sees an underestimated danger in this tendency among soldiers to anthropomorphize machines. In an extreme case, officers could become more attached to their robots than to the men and women they command. Then, so Arkin’s tongue-in-cheek prediction, an officer might well issue an order like, “Tom, you go and see if the coast is clear — the robot stays here!”
Translated from the German by Christopher Sultan