Human Nature

Predators Need Editors

Killer drones aren’t robots. And to protect civilians, let’s keep it that way.

Now that NATO is using unmanned aerial vehicles to destroy Libyan artillery—now that drones are a mainstream global weapon—it’s time to clear up a misconception about them. Drones aren’t robots. They’re remotely piloted by humans. And our challenge is to keep it that way.

An Air Force officer remotely pilots an MQ-9 Reaper aircraft

Critics of the Libya campaign call U.S. drones “death robots,” “flying death robots,” and “flying robots that vaporize people.” Headlines speak of “Robots in the Air” and a “Robot War Over Libya.” And check out the coverage of a new British report on unmanned weapons. The Daily Mail: “Rise of the machines: Unmanned aircraft are moving us towards a ‘Terminator-like’ world, says Ministry of Defence.” The Telegraph: “Drone strikes herald ‘Terminator-like reality’, MoD warns.” The Guardian: “The Terminators: drone strikes prompt MoD to ponder ethics of killer robots.”

Some idiot in Slate called U.S. drone patrols in Pakistan a “robot proxy war” and now frets about “Terminators to Tripoli.” Oh, wait—that’s me.

The robot metaphor implies that drones reduce the human judgment that goes into each firing decision. The opposite is true. Drones increase our ability to scrutinize targets before we hit them. Our job is to exploit that ability.

The British report explains this paradox. It notes

the greater situational awareness provided by the sensors on a persistent unmanned aircraft that observes the battlespace for long, uninterrupted, periods which enables better decision making and more appropriate use of force. This is enhanced by the fact that the decision-maker is in the relatively stress-free environment of an air-conditioned cabin instead of in a fast jet cockpit.

In other words, a drone pilot can think more clearly and at greater length before firing, precisely because he isn’t there.

In today’s Washington Post, Walter Pincus highlights another way in which drones can increase human input. Summarizing a conference on unmanned air power at the International Institute for Strategic Studies, Pincus reports:

Lt. Col. Bruce Black, program manager for the Air Force Predator and Reaper aircraft, noted that some 180 people are involved in each drone mission. The result, he said, is that “there is more ethical oversight involved with unmanned air vehicles than with manned aircraft.” At the same conference, former CIA director Michael V. Hayden described how, with a Predator circling overhead, those involved in ordering use of its missiles from thousands of miles away can call up computer maps that show the potential effects of each weapon. Before any of the Hellfire missiles are launched, he said, the back-up team asks for the “the bug splat” of the attack—a readout of the impact the missile would have on its ground target. Nothing comparable can be done with ground-supporting manned aircraft, he said.

Everything these British and American officials say makes sense: Drones can give us more time, more composure, more eyes on the screen, and more voices in the weapon command and control chain to avoid civilian casualties. And that’s exactly what’s happening in Libya, according to a NATO statement issued Sunday:

NATO destroyed an SA-8 surface-to-air missile yesterday evening at 2040 GMT in Tripoli using the Predator Unmanned Aerial Surveillance weapon system. The operators of the Predator were able to detect a number of civilians playing football near the missile and firing was delayed until the people had dispersed.

But happy stories like this one assume we’ll use drones to minimize mistakes. We might not. We might use them instead to replace human oversight. We might treat them, in short, as robots.

Britain’s report says its defense ministry is  “looking to increase levels of automation,” in part “to make systems more effective in performing increasingly complex tasks,” and in part “to make manpower savings—enabling one operator to oversee a number of unmanned systems simultaneously.” The report notes:

An unmanned aircraft with an automated control system that is designed to reduce pilot workload, so that it is monitored rather than directly controlled, may well react rapidly to self-generated inputs, but the remote operator will be less aware of what the platform is doing on a real-time basis.

This is the real danger of drones: that we’ll turn them into labor-saving devices. We’ll take people out of the weapon control chain because they’re expensive. We’ll trade civilian safety for cost control.

But that tragedy hasn’t happened yet. And nothing about drones makes it inevitable.  The menace at the heart of the unmanned weapons revolution isn’t robots. It’s us.

(Readings I recommend: David Ignatius opposes drones in Libya because they’ve “ become for many Muslims a symbol of the arrogance of U.S. power.” Glenn Greenwald says if we use drones to target Libya’s leaders, we’re violating Obama’s promise not to seek regime change. Kenneth Anderson at the Volokh Conspiracy thinks the drones will be used to distinguish enemy combatants from civilians—and can do this better than jet pilots can. Steven Metz argues that drones will serve NATO’s initial mission in Libya, not expand it. Adam Serwer at the Plum Line says drones will probably kill fewer civilians, but they signify “ a long, open-ended involvement in Libya with no foreseeable endpoint.” James Fallows doubts the U.S. has a plan beyond the drones.)

Like Slate on Facebook. Follow us on Twitter. Human Nature’s latest short takes on the news, via Twitter: