On July 31, the Center for a New American Security published two papers from a project dedicated to understanding how 16 years of extensive drone use have affected the dynamics of national security decisionmaking, particularly regarding the use of force, and exploring the ways those dynamics may affect democratic accountability, congressional oversight and democratic control over the use of force. Behind the Magical Thinking examines the legacy of the Obama administration’s relationship with drones based on extensive interviews with senior officials. This essay is an excerpt from the final publication.


Americans are, generally, positive about drones as a platform – which is a strange measure to have at all but indicative of the degree to which drones have captured public imagination. That drones are frequently described as precise instruments of warfare, carrying out surgical strikes while reducing risks to American forces, is an element of this approval and fascination. That the Obama-era policy on strikes outside the area of active hostilities required “near-certainty” – the highest possible standard – that no civilians were in the vicinity lends the platform a sense of righteousness. But with such standards come high expectations. Sarah Kreps, in one of her studies on public opinion and drones, found: “Domestic US public support for lethal drone operations is high, but that support declines greatly when individuals are informed of civilian casualties.” James Igoe Walsh, in another survey experiment, noted that calling out the precision capabilities of such weaponry actually makes people more sensitive to civilian harm.

Lofty rhetoric – for years backed by little official data – may have given the American people high and not necessarily substantiated expectations of drone platforms as a more “humanitarian” approach to countering national security threats. Indeed, Julia MacDonald and Jacqueline Schneider found gaps in understanding of drones’ true capabilities in another survey. In contrast, many advocacy groups have worked tirelessly to document their sense of noncombatants killed on acknowledged and unacknowledged battlefields – assessments that, in their view, stood in contrast to formal government views on drone precision, even as lethal strikes became more public in the latter half of the Obama administration.

That limited public data and lack of informed debate may have skewed public views of drones more positively is plausible. But a key question of this study is what policymakers’ own views of drones – particularly as they relate to civilian casualties – reflect, and how those views affect policymaking. Unquestionably, for many years policymakers had access to better data on drone strikes and drone platform capabilities than the public did. But that there was no simple way to evaluate the data or constructively challenge it was limiting, both for how policymakers viewed it themselves and how they viewed critics.

A clear takeaway from the interviews conducted in the course of this study is that many Obama administration policymakers did worry, a great deal, about the potential for civilian casualties via drones (or any other kinetic capability). They believed that the policies and practices they shaped were serious efforts to prevent or limit civilian casualties to the greatest extent possible, and on the occasions casualties did occur, they made efforts to learn from mistakes. To be clear, their worries were not due to drones’ propensity for “causing” civilian casualties. There was a general sense among respondents that drone capabilities such as persistence, advanced sensors, and munitions give them unique strengths, that these capabilities were improving over time, and when employed correctly, drones affect civilians at no higher a rate, and almost certainly at a lower rate, than other common means of warfare. Policymakers were not alone in this belief. The Stimson Center notes in the report of its Task Force on US Drone Policy, regarding the end of the Obama administration, “The frequency and number of civilian casualties resulting from US drone strikes also appear to have dropped sharply in recent years, as UAV technologies have improved and targeting rules have been tightened.” Generally, former officials who discussed it believed the U.S. government had a better story to tell than allegations of systemic underreporting of civilian casualties or overestimating of precision strike capabilities would imply.

The Obama administration codified these views with its issuance of the executive order on civilian casualties. The policy generally served as a means to set on paper practices the administration was already pursuing, but it also required a “Report on Strikes Undertaken by the U.S. Government Against Terrorist Targets Outside Areas of Active Hostilities.” This was an important signal on the seriousness of the matter to that administration. The Trump administration is now required by Congress to publish a similar report.

However, some project respondents expressed greater worry around three related matters that have less to do with the precision of drones’ capability and more to do with foundational analysis and process: first, how closely civilian policymakers are sometimes involved in the review of intelligence that identifies targets and clears drone strikes, including what constitutes a continuing imminent threat (while potentially a positive in terms of their attention, also possibly a negative in terms of division of labor and setting clear standards); second, how constrained policymakers have been over time in discussing and receiving feedback on drone strikes publicly; and third, how disconnected policymakers necessarily are from the aftermath of such strikes, particularly outside the areas of hostilities. Policymakers want to understand more about drone strikes, how they are perceived, and their effects, but current practices limit their capacity to do so. Changing this dynamic would require shifting priorities.

These concerns underscore the importance of the underlying intelligence and analytic methods that enable the lethal drone program and review its effects. According to both this project’s interviews and public reporting, policymakers clearly understood that the drone program was only as good as the intelligence that supported it. And during its time in Iraq and Afghanistan, the United States learned many hard-won lessons on what is necessary to protect and support populations that are not easily divided from adversaries.

But some analysts are worried that because of time, rotation, and clear bias against boots on the ground, the U.S. government has unlearned lessons and drifted away from best practices in protecting civilians, and as a result may settle for less when engaging in light-footprint warfare. “Intelligence” one respondent said, “is a demand-driven enterprise; ask better questions and you get better answers.” And with a smaller presence on the ground, the basis for better questions is limited. Moreover, the “cult of secrecy” around drones, as one respondent put it, and disagreements on governmental and nongovermental after-action analysis, may have stunted a necessary public discussion on the realities of how light-footprint warfare is constrained in its discernment and effects. To the extent that such discussion is ongoing, it is both limited and, often, adversarial.

Consequently, there may be a substantial divide in the ways policymakers, the military, and the public think about the impact and cost of drone operations, and the reality on the ground. Air Force officers Colonel Mike Pietrucha and Major Mike Benitez have written eloquently around the dangers of magical thinking around “precise” application of airpower:

Unrealistic expectations surrounding the application of force are making the strategic utility of precision far less than it ought to be – ultimately hindering both strategy and operational utility of the U.S. military. The ubiquitous nature of precision has resulted in the growth of a generation of policymakers who misunderstand the nature of warfare. … The allure of precision weapons has proven too much for policymakers. They have been seduced into believing that somehow, aerial warfare is not the dirty, dangerous, and destructive child of modern warfare that it actually is.

This may be an overstatement, but it is a useful warning about how senior officials discuss warfare in public, how Americans hear them, and potential gaps in their perceptions. Whether or not policymakers view light-footprint warfare as so clean and constrained in nature, they certainly discuss it that way. John Brennan, former director of the CIA, applied such language: “We’re trying to be as careful as a surgeon’s scalpel in terms of taking out the cancer of these terrorist organizations. We have to make sure, though, that we’re not going to damage the surrounding tissue.” Though standards are different there, senior officials in both the current administration and the previous one have called the air campaign against the Islamic State in Iraq and Syria “the most precise in history” – and they mean it. When military officials have been asked about what appeared to be increasing civilian casualties in the fight for Mosul and Raqqa, they have seemed confident that their measures were appropriate. “Civilian casualties are a fact of life in this sort of situation. … We do everything humanly possible consistent with military necessity, taking many chances to avoid civilian casualties at all costs,” Secretary of Defense James Mattis commented last spring. Lieutenant General Steve Townsend has been particularly emphatic that coalition forces are highly restrained, accusing human rights organizations of supporting policies (like operational pauses) that would encourage ISIS’s use of human shields.

But the U.S. government is more ­– or at least differently ­– constrained in its ability to glean intelligence to identify targets and actually conduct post-strike battle damage assessments, particularly as compared with what it might have done in prior years in Iraq and Afghanistan. The “by, with, and through” approach puts these burdens largely on partner forces or air-based ISR. With fewer, or in some cases, no ground groups available to conduct investigations – or with such missions being lower priority than counterterrorism missions – policymakers are comparatively limited in how they can confront the ramifications of drone strikes. Nor are nongovernmental organizations (NGOs) or media with greater on-the-ground access always a trusted source for credible reports of civilian harm, and no agreed-on methodology exists among government and nongovernment stakeholders to bridge the gaps in their assessments.

As a result, some policymakers interviewed for this study were frustrated by the ongoing back and forth between the military and the NGO community, believing both that NGOs to be critical partners in the mission of civilian protection and that “we have a better story than people are imputing to us” but are poorly suited to telling it. Still, credible external reports, such as the recent New York Times investigation “The Uncounted,” have contradicted this narrative, criticizing how the U.S. selects targets, how precise its weapons actually are, how it assesses accusations of civilians casualties, and how it engages known victims. Furthermore, a great deal of work has been conducted in this regard by the Bureau of Investigative Journalism, the New America Foundation, FDD’s Long War Journal, Airwars, and others, to catalog strikes and account for affected civilians using their own methodologies, which the U.S. government disputes. The resulting lack of trust between such organizations and the government makes it challenging to engage on matters of civilian harm in a healthy way.

Clearly, policymakers, military officials, and civil society groups are not approaching the matter of precision and civilian casualties in the same way. It’s reasonable to give the Defense Department credit that its operations are precise by comparison with historical examples, and the department does take more measures than any allies or partners to protect against civilian harm. But it also seems reasonable to theorize that even if the government and human rights groups agreed on the size of the impact of air strikes on civilians on a given battlefield, as well as the capabilities of precision weaponry, they would still disagree on the nature of harm to civilians and its effect on U.S. national security interests. That the United States is far ahead of others in caring about civilian harm does not exempt it from examining its own assumptions about how to measure and evaluate it. Presently, the government does not demonstrably take into account drone blowback that “can anger whole communities, increase anti-US sentiment and become a potent recruiting tool for terrorist organizations.” Nor does it publicly measure the fear of those who aren’t targeted, the physical impact of munitions on communities even when the right targets are hit, or the international perceptions of drone strikes as excessive. And even where it does measure harms to noncombatants, increasingly, it is slow, if not at a standstill, in processing grievance payments. The human factor is missing.

There should be no surprise that, after nearly two decades of ground-troop-intensive warfare, the evolution to light-footprint approaches would require a reset on U.S. approaches to civilian harm. Such a reset demands refinement to both tactical approaches and strategic understanding. Lessons on civilian harm learned well in earlier stages in Afghanistan need to be revisited and refined for air-centric approaches. For example, Rogers, Reid, and Kolenda,’s report “The Strategic Costs of Civilian Harm” highlights a study by Dr. Larry Lewis finding that

air [battle damage assessments] can have substantial blind spots, particularly in the absence of other sources of intelligence and on-the-ground information, and may not be reliable on their own as a means of assessing civilian harm. According to one study in Afghanistan, initial air BDAs failed to identify civilian casualties in 19 out of 21 [italics in the original] cases subsequently confirmed by ground force investigations.

If the United States is unable to deploy necessary ground troops to make up that difference, it should be willing to consider alternate means to both enhance the legitimacy of target selection and measure and respond to civilian harm. Starting that conversation does not minimize the significant efforts already undertaken by the U.S. government to protect civilians; rather, it recognizes the changing nature of the battlefield. But more broadly, as Rogers et al. note, senior policymakers must grapple with the fact that in these irregular conflicts, “civilian harm, even in accordance with [the Law of Armed Conflict], [italics in the original] can cause irreversible damage to a U.S. mission – a serious risk that also applies to U.S. counter-terrorism operations and partnerships with foreign security forces.” It is here that senior policymakers and military officials must both internally and externally get into the weeds of definitional and risk debates in a way that will almost certainly be uncomfortable – but fruitful.

Image: An MQ-9 Reaper sits in a hangar during a sandstorm at Joint Base Balad, Iraq, Sept. 15, 2008. U.S. Department of Defense.