notes.husk.org. scribblings by Paul Mison.

Andrew Cockburn On Drones

Andrew Cockburn’s recent subscriber-only article, Drones, Baby, Drones in the London Review Of Books covered the US military’s recent favourite toy, its history, and why it may not be all it seems. I thought it deserved a summary for a wider audience.

The US has tried automated systems before:

Igloo White, which cost $7 billion, was an early attempt to automate the battlefield: tens of thousands of sensors, designed to pick up sound or movement, each one in radio contact with computers in Thailand, were scattered around the jungles of Vietnam and Laos in the hope of locating and targeting enemy supply columns on the Ho Chi Minh trail. But the Vietnamese quickly learned to move the sensors or make them send false signals and the experiment was abandoned in 1972.

There were also much-publicised claims for advances in technology in the Gulf War, the bombing of Serbia, and in Iraq after 2003. In each case, watchdogs (such as the US Government Accounting Office) claimed after the conflicts - when nobody was watching - that the programmes tended to produce much worse results than were claimed.

The last decade’s conflicts in Iraq and Afghanistan, and the distaste for casualties from the western public, have meant a bonanza for those making automated systems:

While American and British casualties on the ground in Iraq and Afghanistan ticked upwards, the skies grew dark with target-seeking aircraft, manned and unmanned, operating under such codenames as Constant Hawk and Angel Fire. Many of these systems were vastly expensive: one airborne system for detecting roadside bombs, Compass Call, cost $70,000 an hour.

Unfortunately, as with previous systems, they may not have actually worked that well:

In 2007, an intelligence unit in Baghdad called the Counter-IED Operations Integration Center analysed hundreds of these missions and concluded that most of them had ‘no detectable effect’ on the enemy.

That hasn’t stopped the funding for new projects, such as Blue Devils and Gorgon Stare, developed by MAV 6, whose CEO, David Deptula, used to be an Air Force three-star general. That said, maybe we don’t have much to fear:

Gorgon Stare’s camera images could not distinguish humans from bushes, or one vehicle from another. It had severe problems working out where it was. It broke down, on average, 3.7 times per sortie. The testing unit recommended that it shouldn’t be deployed, advice rejected by higher authorities, who quickly dispatched it to Afghanistan.

It’s not just the military. Despite Rick Perry’s pledge during his abortive run to be a presidential candidate that he’d deploy drones at the Mexican border,

when the Department of Homeland Security compared the relative performance of unmanned Predators and manned light Cessna aircraft in detecting illegal border crossings, it found the old-fashioned plane to be ten times as effective and 30 times cheaper. (Keeping a drone in the air involves two or three times as much backup manpower as a jet fighter like the F-16.)

It’s bad enough that drones are expensive and unreliable, but as events in Iran have shown, they’re also vulnerable to hacking:

In a less publicised incident last March, an American reconnaissance plane over South Korea suddenly returned to base on discovering that its GPS was being jammed, an imprudent move since it confirmed to the North Koreans the effectiveness of their methods.

Sadly, as the article concludes, the Air Force and its suppliers are tied together with lobbyists, and whether or not their unmanned aircraft actually work seems to be beside the point. Deptula argues now that the problem is that there’s still a “man in the loop”, the remote pilot who actually flies the drones (which, for all the hype, are still far from automated). This seems mildly insane to me, but then, the future will no doubt show which of us is right - but wait for the audit commission’s report rather than believing CNN.

(I very much enjoy my subscription to the LRB, despite being perpetually an issue behind in the physical edition. I’d heartily recommend one if you want to read the full post.)