Killer robots, part 2: implications for international law?
As Foreign Policy Watch recently noted, the winter edition of the Wilson Quarterly features an article by P.W. Singer on the vertigo-inducing growth of robots and automation in the U.S. military, adapted from his new book Wired for War.
Opinio Juris has a nice summary of points raised during an NPR interview with Singer, particularly his theme that the assumptions by which we define “war” are less and less valid. This is critical, and it seemed to be a point overlooked in a Washington Post opinion piece that caromed around a few blogs recently.
The literature that’s sprung up so far seems to be unified by the contrast between the efficacy of military robots, and the ethics of their use. The legal question, when examined, seems to concentrate on whether we can avoid the robot equivalent of a My Lai (and, indeed, John Pike, among recent authors, has even suggested robots could thwart genocide, not to mention lesser atrocities), given the potential lethality of autonomous or semi-autonomous machines.
While important, I’m not sure that the efficacy/ethics question can be resolved without a better understanding of what now constitutes “war,” as Singer’s work suggests. Take for instance, the example cited of the human operator in the U.S. piloting a drone operating in a war zone a continent away. Singer elsewhere in the article quotes Gordon Johnson of the Joint Forces Command: “The enemy, are they going to give up blood and guts to kill machines?” Left unanswered is the question as to whether an enemy would instead focus on attacking the man-in-the-loop in the U.S., and whether or not this would constitute a legitimate act of war. Is the man-in-the-loop a combatant in this instance? Will increased use of robotics, controlled from afar, not just change the proclivity to use military force, as Singer asks, but also blur the boundaries of what legitimately constitutes a “battlefield”?
As Gary M. Anderson and Adam Gifford, Jr., have noted, international law has historically lagged technological change – the submarine and strategic bombing representing two cases in point. But in a situation where Moore’s Law is driving the rate of change – and where advances in robotics are not limited to the U.S. – would it not be wise for a post-Iraq U.S. to lead a definition of conventions for robot operation and use? Leaving this one to technology evolution just seems too dicey. I don’t know whether this gets raised in Singer’s book, but I’m looking forward to reading it.