Armed drones may be just the beginning of a major technological revolution in the defense industry. An expert on military robotics, P. W. Singer says the “Predator” and “Reaper” drones flying over Pakistan and elsewhere are “the equivalent of the Model T Ford or the Wright Brother’s Flyer.” Indeed, more can and is already being done in the military robotics industry…lots more. The Washington Post ran a story yesterday on efforts by U.S. engineers to develop fully autonomous robotic weapons, that is robots that could “hunt identify and kill the enemy based on calculations made by software, not decisions made by humans.” In other words, imagine the Terminator.
Some might think this is a bit too outlandish. I suggest you read P. W. Singer’s book, “Wired for War,” a well – sourced and extremely accessible book on the military robotics industry. Singer, a leading expert on this issue, suggests it’s really only a matter of time before “killer robots” become a reality. Advanced militaries, including the U.S., U.K., Israel and South Korea, are spending billions of dollars every year on drones and other types of autonomous robots. Some of the weapons developed by these states are very close to what one might call a fully autonomous weapon. Earlier this year, the U.K. Ministry of Defense suggested that fully autonomous weapons could exist within 5 – 15 years. However, most engineers believe it will be more than fifteen years before fully autonomous weapons are introduced to combat.
The problem is that a very sophisticated level of “artificial intelligence” would be needed to ensure such weapons comply with the laws of war. The most difficult norm for autonomous robots to observe would be the principle of distinction, which requires that belligerents distinguish between combatants and noncombatants in the use of force. Human soldiers are having more and more trouble abiding by this norm in contemporary conflicts because insurgents often hide amongst the civilian population. Identifying combatants in these conflicts requires very sophisticated levels of intelligence and human reason, which engineers may not ever be able to replicate in robots.
There are lots of legal and ethical problems with killer robots. Most problematic, however, is that despite a fairly significant amount of literature on this subject, there is surprisingly little policy debate on whether fully autonomous killing machines are a good idea. Perhaps, some believe it is too absurd or far off to merit attention. Those that are skeptical should just listen to what the engineers working on these weapons say. Ronald Arkin, an engineer funded by the U.S. army to research this issue, believes “lethal autonomy is inevitable.” That is a scary thought indeed.
