A slew of news reports have highlighted the crisis of drone pilot burnout in the United States military. Indeed, pilot shortages have prompted the US Air Force to cut the number of drone flights to fewer than 60 per day. That’s an important problem, but buried in these stories is another one. The Air Force has announced that, in response to the shortage, it will increase its use of contractors for these flights. Given the service’s manpower shortages, this statement is not surprising. Yet the growing numbers of contractors in drone operations, while little discussed, raise significant concerns about oversight and accountability at a time when drone use is set to accelerate. We simply don’t know enough about how contractors will be used in the increasingly automated version of war that appears to be our future. And that means we need to ask hard questions now about how this system should operate rather than simply letting it evolve without oversight.

Many people don’t realize that it actually takes dozens to hundreds of people to run just one mission of the larger drones. For example, a single 24-hour combat air patrol mission for Predator or Reaper drones requires approximately 160 to 180 people. One mission of the more complex Global Hawk surveillance drone demands between 300 and 500.

The Air Force has repeatedly said that only uniformed personnel actually fly drones that carry bombs and engage in targeting. Recent Air Force statements about the increased role for contractors do not anticipate a major shift in this policy, (although officials have said that contractors could perform limited flying roles such as assisting with takeoff and landing). But even in supporting roles, contractors’ actions could raise concerns. A study from 2012 estimates that contractors fill 75 to 100 percent of maintenance jobs for some categories of drones and 10 percent of jobs related to intelligence processing, exploitation, and dissemination.

And while the Air Force has so far not allowed contractors to operate drones that carry bombs, contractors do operate smaller Air Force surveillance drones, like the Scan Eagle and the Raven. The pilots of these drones often work in theater, controlling the vehicles from the ground and collecting intelligence information that then serves as the basis for targeting decisions.

To see why even this level of contractor involvement could be a problem, consider a 2010 incident in the Oruzgan province of Afghanistan. Hellfire missiles launched from an Air Force Predator killed 15 Afghan civilians and injured a dozen more traveling near US special operations forces who were conducting a capture mission. Subsequent investigations revealed that, although military personnel were operating the drone and the ground force commander made the decision to strike, the decision to fire was largely based upon an intelligence analysis that a civilian contractor had provided.

We know much less about the operation of the CIA’s drone program, but reports have indicated significant contractor involvement there as well. For example, contractors have assembled the bombs that get loaded onto the aircraft. In some cases, these bombs have not exploded, raising questions about contractor performance. In another reported case, a bomb assembled by the contractor firm formerly known as Blackwater fell off a drone before it reached its target, leading to a search for the unexploded weapon.

While we don’t know nearly enough about how contractors are and will be used in drone warfare, we can look to the past role of contractors in Iraq, Afghanistan, and the so-called “War on Terror.” In each, a toxic brew of anemic governmental oversight, inter-agency squabbling and deception, and a culture of impunity made it possible for contractors at times to send governmental and military authorities down a path of poor judgment, based on little experience, to misguided adventures in abuse. The examples are plenty: The recent revelations of the Senate Intelligence Committee report on the CIA torture program, which showed how contract interrogators led the charge for harsher and harsher interrogation tactics such as waterboarding; the role of contractors who reportedly supervised military interrogators while they abused detainees at Abu Ghraib Prison in Iraq; the notorious incident in Baghdad’s Nisour Square in 2007, when Blackwater security guards under contract with the State Department, fired into a crowd, killing 17 civilians. All are instances of contractors at the forefront of abuse.

More frequently, the hundreds of thousands of contractors in Iraq and Afghanistan were just doing their jobs, not participating in abuse. And thousands died serving US interests. But the explosion in the number of contractors, combined with reduced funding for contract monitors, resulted in more than $30 billion in waste, fraud, and abuse, according to the Commission for Wartime Contracting.

Ever since many, including myself, first pointed out these problems over a decade ago, some progress has been made. Indeed, my 2011 book on the subject, Outsourcing War and Peace, proposed numerous reforms, and since then Congress, government agencies, and the contracting industry itself have pursued many of them. But have the reform efforts gone far enough? How prepared are we for the next deployments of contractors operating drones and conducting other high-tech military operations? We can look at two areas of concern: oversight and accountability.

First, lack of transparency about the role of drone contractors hinders effective oversight.  The fact that some contractors perform imagery analysis that may contribute to targeting decisions is not a problem in and of itself.   But it becomes problematic if the contractors are effectively making the targeting decisions.  The contracts’ terms may forbid them from doing so, and may require governmental personnel to supervise the contractors.  But if the ratio of contractors to governmental personnel grows too large, the supervision ban could be rendered meaningless.  This is particularly true when the contractors are seen as having expertise or experience that the uniformed military personnel lack, leading to an undue reliance on the contractors’ interpretations and recommendations.  That’s why at the very least the Defense Department ought to report on the breakdown and roles of contractors involved in operating drones as part of its quarterly report on contracting.  The public shouldn’t have to rely on piecemeal investigative journalism to get a glimpse of the contractor function here. And that’s also why, especially now that mandate of the Special Inspector General for Iraq Reconstruction (SIGIR) has lapsed, we need a permanent inspector general to monitor contractor activity.

Second, if a drone contractor were to commit a crime, accountability remains a big issue, particularly for CIA contractors operating overseas. The very jurisdiction of U.S. courts to consider such cases remains unclear because the primary applicable statute, the Military Extraterritorial Jurisdiction Act, only applies crimes committed by DOD contractors or those whose employment relates to a DOD mission.  Thus, contractors such as the State Department security contractors implicated at Nisour square or CIA contractors operating drones overseas do not easily fall within US courts’ jurisdiction.  The proposed Civilian Extraterritorial Jurisdiction Act, which would close this loophole and clarify that contractors for all agencies could be subject to prosecution, has been languishing on Capitol Hill for years.  The bill would also provide for better inter-agency coordination of investigations.  Indeed, even many in the industry support it because contract firms have realized it’s actually good for business to ensure that rogue actors are punished.

In addition to criminal punishment, the tort system also plays an important role. Yet the legal framework here is in perhaps even more disarray than on the criminal side. Some courts, such as the DC Circuit, have taken an overly broad view of an immunity doctrine that shields contractors from tort liability for battlefield conduct. And the US Supreme Court has interpreted the Alien Tort Statute very narrowly, making it difficult for non-US citizens to sue US contractor firms in US courts. Moreover, with regard to autonomous weapons, the opaque web of decision-making responsibility will likely make it increasingly difficult to trace accountability to specific actors. As such, though the tort system should provide some accountability for contractor abuse, significant legislative reform is necessary to fulfill that promise.

*                      *                      *

We clearly have a long way to go. And as we enter a new round of wartime contracting, immediate action is crucial. Otherwise, 10 years from now we may be dealing with the fallout from another horrific government report, this time dealing with autonomous weapons, errors, violations of the laws of war, targeting decisions made outside the chain of command, and so on. And the answer cannot be simply to oppose the use of contractors: such a position is probably unrealistic. Instead, we need to build structures of oversight, transparency, and accountability into our outsourcing regime. And the rise of complex autonomous weapons systems only makes the need more urgent.