Written by David Cole for The New York Review of Books and RSN

The President of the United States can order the killing of US citizens, far from any battlefield, without charges, a trial, or any form of advance judicial approval. That’s what Attorney General Eric Holder told a group of students at Northwestern Law School yesterday, in a much anticipated speech. The Constitution requires the government to obtain a judicial warrant based on probable cause before it can search your backpack or attach a GPS tracking device to your car, but not, according to Holder, before it kills you.

Holder’s speech marks a victory of sorts for those who have condemned the secrecy surrounding the administration’s aggressive targeted killing program. At a minimum, we now have a better basis for a debate about the extent to which a democratically elected leader should be entitled to single-handedly order the execution of those he represents. So those inside the Obama administration—including State Department Legal Adviser Harold Koh—who reportedly fought a pitched battle for this disclosure, deserve credit for the increased transparency it has brought.

But on the merits, the executive authority Holder asserted is deeply disturbing in the days of lethal strikes by unmanned drones. Garry Wills argued in Bomb Power that the nature of the Presidency was fundamentally altered with the introduction of the nuclear bomb; but in some ways, drones may ultimately mark an even more tectonic change. The nuclear bomb is so devastating that it cannot realistically be deployed (and has not been used since we dropped them on Hiroshima and Nagasaki in World War II, killing more than 200,000 people). The drone, by contrast, can be deployed, and has been, with increasing frequency. It allows for relatively pinpoint targeting, and the collection of detailed intelligence about suspects’ whereabouts so that, at least in theory, collateral damage can be limited. And perhaps most significant of all, those directing the drone—from computer screens at military bases thousands of miles away—face no risk of loss of life. Thus, the built-in check on killing, namely, that the one engaged in the killing risks being killed himself, is gone. Drones offer a new and seemingly costless (apart from the $4.5 million price tag) way of doing battle, and therefore change the calculus of war dramatically.

The extent of the change is reflected by the fact that no president has previously asserted the power to order the killing of an American citizen far from the battlefield. If you are inclined to trust Obama with such power, what about the next administration? Or the leaders of Saudi Arabia, Russia, or China? In international law, what the United States does is often a precedent (or pretext) for others, and we will not have a monopoly on drone killing for long.

So how does President Obama, the constitutional law professor who has vowed to fight terrorism within the constraints of both domestic and international law, justify such a dramatic taking of life without judicial process? It is not illegal or even controversial, of course, to shoot to kill enemy soldiers on a battlefield in wartime. An American citizen who chooses to fight for the other side takes the risk that he will be targeted along with his fellow enemy soldiers.

But while he didn’t mention him by name, Holder was out to justify the killing in September 2011 of US citizen Anwar al-Awlaki, an American preacher and propagandist living in Yemen, more than 1,500 miles from the Afghan battlefield. Al-Awlaki was allegedly connected with the Christmas Day “underwear bomber” Umar Farouk Abdulmutallab and a foiled plot to place bombs on cargo planes flying from Yemen to Chicago. But while the underwear bomber was arrested, charged, and convicted by a US federal court, al-Awlaki was simply eliminated with a drone strike.

Holder asserted that the president can order a targeted killing when: (1) we are at war; (2) the target lives abroad and is an operational leader of al-Qaeda or an “associated force”; (3) there is no feasible option for capture; (4) the individual poses an “imminent” threat of attack; and (5) the order is carried out consistent with law-of-war principles governing the use of force. Killing even a US citizen in those circumstances is consistent with “due process of law,” Holder asserts, even if no court reviews the executive’s decision before or after.

But each of the factors Holder lays out raises as many questions as it answers. If the “armed conflict” with al-Qaeda has no end in sight, is this effectively a permanent standing authority in terrorism cases? Will the authority continue to exist if and when we pull out of Afghanistan? (The Administration seems to suggest that it will, in view of the fact that al-Awlaki was not in Afghanistan or directly connected to that conflict.)

Second, what constitutes an “associated force?” Al-Awlaki was said to be a leader of al-Qaeda in the Arabian Peninsula, or AQAP, but that entity did not even exist on September 11, 2001, and was created only in 2009. Must there be evidence that the associated force has coordinated military activities against the United States in connection with the ongoing armed conflict in Afghanistan, or is any terrorist group inspired by Osama bin Laden’s rhetoric an “associated force?” Holder does not say.

Third, what does it mean to say that capture is not feasible? Holder says only that it’s a “fact-specific” inquiry that “may depend on, among other things, whether capture can be accomplished in the window of time available to prevent an attack and without undue risk to civilians or to US personnel.” In al-Awlaki’s case, he was put on the targeting list in early 2010, so there was a pretty long “window of time available.” Reportedly Yemen tried to capture him once, but he escaped. Does one failed attempt make capture not feasible? And perhaps more disturbingly, does the fact that drones make it possible to kill without any risk to US personnel make any risk to US personnel “undue”?

Fourth, and in many ways most problematically, what constitutes an “imminent” threat of violent attack? Much has been leaked about al-Awlaki’s alleged involvement in terrorism, but no one has claimed he was involved in any particular attack at the time the administration killed him. How can an attack be “imminent” if no attack is about to be launched? According to Holder, “the evaluation of whether an individual presents an ‘imminent threat’ incorporates considerations of the relevant window of opportunity to act, the possible harm that missing the window would cause to civilians, and the likelihood of heading off future disastrous attacks against the United States.” But he also claimed that al-Qaeda is “continually planning attacks” and “has demonstrated the ability to strike with little or no notice,” so therefore the President need not wait until “the precise time, place, and manner of an attack become clear.”

This appears to define “imminence” away. If al-Qaeda and its associated forces always present what the administration defines as an “imminent” threat then imminence ceases to have meaning. But this requirement is critical, because surely such killing, away from the battlefield, should only be undertaken as a last resort. As long as the individual is not engaged in an imminent attack, there is always a possibility that his capture may become feasible. Imminence is designed to ensure that lethal force is a last resort. But Holder and the Obama administration appear to have turned it into just another policy option.

Finally, does due process require some sort of judicial process? Holder said no: “The Constitution guarantees due process, not judicial process.” But in this setting, why wouldn’t judicial process be required? Holder admitted that due process applies, at least when a citizen is targeted. The Supreme Court has long said that due process requires balancing the private interest at stake, the governmental interest, the risk of error, and the burden of providing more process. Holder conceded that the interest in not being killed illegally is paramount, but noted that so is the government interest in preventing an imminent attack.

But he paid far too little attention to the risk of error or the burden of providing further procedural guarantees. The risk of error where the executive acts as prosecutor, judge, jury, and executioner, in secret, could hardly be greater. One need only look at the over 600 men once condemned as the “worst of the worst” but now released from Guantanamo to see that the executive can make mistakes. Holder provides no assurance that a demanding review process is undertaken, or of how it is constructed to minimize risk of error, or even what standard of proof is employed.

Moreover, he opposes any judicial process, even where there is time to provide it. In a case like al-Awlaki’s, where a person has been on a “kill” list for nearly two years, surely the minimal burden of making one’s case to an independent judge in camera would be worth the time and effort required. Notice and an opportunity to be heard may not be realistic, but Holder did not explain why it would impossible for the executive to make its case to an independent judge, much as it does for “foreign intelligence” searches and wiretaps under the Foreign Intelligence Surveillance Act. We have long recognized that those charged with law enforcement or security are at risk of overestimating their own certainty, and have therefore required that the government obtain a warrant from an independent judge before conducting a search, unless there is not time to do so. If we require such process even for the search of a backpack, shouldn’t we demand at least as much before the President orders the non-battlefield killing of a human being?