Civilian Casualties, Drones, Airstrikes and the Perils of Policy


Last week, the Daily Beast breathlessly reported an “exclusive” story, alleging that Department of Defense officials admitted that anti-ISIS airstrikes had killed what the Beast characterized as “innocents.”

The Beast obviously thought it was on to something big with its “exclusive.” So what number of civilian deaths (out of attacks on more than six thousand targets) did the Beast unearth? Two. As if recognizing weakness of that number as supportive of an indictment of what the Beast calls a “relentless bombing campaign,” the Beast added the allegation by a Syrian organization that 64 civilians were killed in another airstrike.

For the sake of argument, accepting the worst case figures as the Syrian group did (and even adding earlier allegations of “dozens” more from other sources), the numbers still reflect an astonishing — and even unprecedented — precise and discriminate application of force given the thousands of targets struck.

Yet this story, along with a host of stories about civilian casualties allegedly caused by airstrikes from drones, all share something in common: an ill-concealed hostile tone. Moreover, they frequently marginalize the overall military effectiveness of airstrikes (if it is mentioned at all). There are a lot of reasons for this — including the military’s overselling of precision capabilities — but a very significant one could be the confusion that has arisen about the law vis-á-vis civilian harm.

For the United States, much of this confusion is a self-inflicted wound traceable to the Obama administration’s attempt to explain its use of force approach, and especially its procedures to protect civilians. That effort is laid out in a document entitled U.S. Policy Standards and Procedures for the Use of Force in Counterterrorism Operations Outside the United States and Areas of Active Hostilities, issued in May of 2013.

The unhappy reality is that despite what may have been the administration’s best intentions to demonstrate the “transparency” that so many demand, the memorandum did nothing to silence or even temper the critics, and may have actually endangered the very civilians it sought to protect. The Policy Standards are an object lesson of the perils of trying to appease critics by attempting to use policy to “improve” upon what international law provides and permits.

What the Policy Standards do is brew together bits and pieces of international humanitarian law (the law of armed conflict), international human rights law (essentially law enforcement rules), and political preferences. The result is a formula virtually guaranteeing confusion and discontent, and not the clarity and consensus policymakers might have hoped to achieve.

It is unsurprising, for example, that even when the administration tries to explain that the Policy Standards by their own terms were never intended to apply to operations like those against the Islamic State in Iraq and Syria, it can’t “unring” the proverbial bell. Too many people just do not make the fine geographic and other distinctions upon which the Policy Standards are premised. And it was a mistake to have expected them to do so.

Accordingly, it is legitimately “news” to a lot of people when the administration appears to deviate from the policies even in circumstances under which it plainly did not apply. Thus, easily misunderstood headlines like “White House exempts ISIS strikes from civilian casualty guidelines, suggest a sort of “anything goes” approach, despite the fact that these operations still fully comply with international law. Even more importantly, there seems to be little recognition that the original policy went far beyond what the law — and, frankly, good sense — would require.

While there is much that could be discussed about the Policy Standards, allow me to focus on the portion that calls for targeteers to have a “near certainty that non-combatants will not be injured or killed” in any use of force in counterterrorism operations outside the U.S. or “areas of active hostilities.” That is not what the law requires, and for good reason.

With respect to the law of armed conflict, the requirement is set forth in Additional Protocol 1 to the Geneva Conventions (the United States is not a party to the protocol, but this portion is accepted in most quarters as having attained the status of customary international law). The protocol holds that attacks are prohibited if they “may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated” [emphasis mine].

In other words, contrary to what lots of non-expert readers of the administration’s policy might understandably assume, it is actually perfectly legal to execute an otherwise lawful attack notwithstanding the knowledge that, in fact, civilian casualties will inevitably occur.

Put another way, contrary to the implications of the Policy Standards, under international law you can have a “near certainty” (or even absolute certainty) that civilians will be killed — innocent or otherwise (more about “innocents” later) — and still launch the strike. If civilian casualties do occur in an attack, the strike is not necessarily illegal under international law, even if the casualties are not the result of “mistakes” or “errors” as you see so often speculated in the press.

Why is the law this way? If it wasn’t legal to conduct an attack even when you are certain that civilians will be killed, belligerents would be incentivized to surround themselves with civilians in order to create a legal sanctuary from attack. It really is that simple.

Unfortunately, the administration attempted to “improve” upon international law by announcing directives not just prohibiting all strikes that could result in any civilian casualties regardless of the importance of the target, but also by imposing an almost impossibly high “near certainty” standard on targeteers.

Putting aside the ill-advised “no civilian casualty” rule, a “near certainty” standard about any aspect of the targeting process exceeds the reasonableness standard international law sets for commanders as to use of force decisions. It also surpasses the “objectively reasonable” yardstick that in the United States can justify the use of force in peacetime (and which also resonates favorably with international human rights law that some maintain applies to certain drone operations).

Deviating from international law can have tragic, unintended consequences. Consider the recent deaths of an American and an Italian hostage in a drone strike in Pakistan. Although the reasons as to why these hostages were co-located with key al-Qaeda operatives rather than simply guarded by low-level foot soldiers are unknown, is it not plausible (given the administration’s announced policy) that al-Qaeda might have concluded that the presence of civilians — and especially an American — would amount to a perfect defense against the drones they feared so much? Is it not conceivable that they considered these civilians as, in essence, effective shields against a U.S. use of force?

Of course, the use of human shields is a serious violation of international law, but international law is hardly an impediment to al-Qaeda, especially if violating it might ensure policy-induced protection against airstrikes or other uses of force.

“Near certainty of no civilian casualties” also unrealistically raises expectations for an “immaculate” war where only bad guys get hurt. It is no wonder that critics have a field day when civilian casualties do occur, as is wholly predictable when thousands of targets are struck. Misunderstandings about the propriety of the use of force can have a severely deleterious effect on operations vital to national security.

Michael Reisman and Chris Antoniou have warned that in a democracy, beliefs that an operation is “being conducted in an unfair, inhumane, or iniquitous way” can erode the necessary “base of public support…no matter how worthy the political objective.” In short, people need to know that civilian casualties are not, ipso facto, evidence that something illegal or immoral has taken place, or even that a mistake has been made, but you won’t get that from the Policy Standards.

Finally, what about the Daily Beast’s reference to “innocents” being killed? While the hostages tragically killed recently were certainly blameless civilians, international law actually has no requirement that a civilian be morally — or even legally — “innocent” to be entitled to legal protection from attack.

Americans instinctively want to think that all civilians are “innocents” even when logically that may not be true. As the late historian Daniel Boorstin pointed out, Americans suffer from “Myths of Popular Innocence” which he characterized as “a touching … unwillingness to believe ill of human majorities.” Boorstin said it tends to lead Americans to “exonerate the populace from the horrors in which they must have collaborated.” He cited the complicit citizenry of Nazi Germany as an example of such an evil collaboration yet who have been treated as if they were “innocent bystanders” or simply victims of an “unopposable master.”

In reference to the Kosovo war of the late 1990s, New York Times columnist Anthony Lewis similarly argued that Slobodan Milosevic was “not the only author of the war crimes in Kosovo” as “hundreds of thousands more back in Serbia proper were fixed in the belief that the Serbs had done no wrong” despite irrefutable evidence.

In short, while the vast majority of the populations of Nazi Germany and Milosevic’s Serbia would today be considered “civilians” legally protected from being directly targeted, one can hardly say that those among them who embraced the loathsome ideologies are morally innocent. The same can be said of the supporters of terrorist extremism today.

In some or perhaps many contemporary situations it is quite possible that a civilian killed in an attack may not only be morally blameworthy, but also legally culpable of a crime, including material support to terrorism. These supporting activities, while illegal, do not necessarily amount to “direct participation in hostilities,” which is the level of involvement that international law requires to make civilians (as opposed to other non-state actors) properly targetable. Still, such persons are hardly “innocents.”

This is not to suggest that they ought to be kinetically targeted, but rather merely to make the point that one cannot assume — as the Beast seems to do — that all civilians killed in airstrikes are moral or legal “innocents.”

The takeaway from all this is that clarity as to what is — or is not — required by the law is the kind of transparency we really ought to expect and demand from the government. Of course, it is perfectly acceptable to choose to have policies that require more than what the law might demand, but they ought to be enacted only when there is a clear understanding by decision-makers and the public as to how an enemy might manipulate them to produce wholly unintended consequences.

In any case, restrictive policies should not be put in place out of some naïve hope that most of the critics will somehow be satiated and pacified. They won’t be.

It is far better (and vastly more realistic) to scrupulously adhere to the law — which is typically the product of hard experience in actual conflicts — and strike out at those who are responsible for far more civilian deaths than drones or airstrikes. We need to be mindful that a policy (as opposed to a law) driven decision to forego a strike can produce terrible consequences.

The plain truth is that an overly restrictive targeting policy that delays or even prevents attacking terrorists means that they will live on to continue to wreak havoc against civilians. Who, we might ask, should share accountability for these needless deaths?


Charles Dunlap is currently the Executive Director of the Center on Law, Ethics and National Security at Duke Law School. He retired from the Air Force in 2010 as a major general after more than 34 years’ service.

Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; bingbot/2.0; + Chrome/116.0.1938.76 Safari/537.36