Unmanned aviation systems, popularly known as drones, are playing an increased role in armed conflicts.1 They are used both for collecting intelligence and for deploying lethal force. In 2007 there were 74 U.S. drone strikes in Afghanistan.2 That year, there were five strikes in Pakistan.3 By 2012, the American military was executing an average of 33 drone strikes per month in Afghanistan, and the total number in Pakistan has now surpassed 330.4 Recently the United States has proposed further expanding its deployment of drones, developing plans to set up additional Predator drone bases in Africa that would allow these drones to cover much of the Saharan region.
Drones have been employed in multiple theaters of the counterterrorism campaign, including Yemen, Somalia, Iraq, and Libya. They are now included in the arsenal of many nations including Israel, China, and Iran. They have even been operated by a non-state actor, Hezbollah, which has flown at least two drones over Israel.6 Several nations are currently developing drones that will be able to carry out highly-specialized missions, for instance tiny drones able to enter constricted areas through narrow passages. If the American military continues to move away from deploying conventional forces on the ground (in Iraq and Afghanistan) to a "light footprint" strategy of "offshore balancing" (as employed in Libya), drones are likely to play an even more important role in future armed conflicts. Like other new armaments (e.g., long-range cruise missiles and high-altitude carpet bombing) the growing use of drones has triggered a considerable debate over the moral and legal grounds on which they are used. This debate is next reviewed.
Excessive Collateral Damage?
Critics argue that a large number of civilians, including women and children, are killed by drones. Some hold that the number of civilians killed amounts to an overwhelming majority of all those killed. Syed Munawar Hasan, who heads the influential Islamic political party Jamaat-e-Islami in Pakistan, has claimed that the drone strikes "are killing nearly 100 percent innocent people."7 Former military officers David Kilcullen and Andrew Exum argued in the New York Times that in Pakistan drones kill 50 civilians for every militant. Other critics put forward somewhat lower numbers. A study conducted by the Columbia Law School estimates that 35 percent of the victims of drone strikes in 2011 were civilians. In contrast, American counterterrorism officials put the number as low as 2.5 percent. Deputy National Security Advisor for Homeland Security and Counterterrorism John Brennan claimed that "there hasn't been a single collateral death because of the exceptional proficiency, precision of the capabilities we've been able to develop."8
Researchers who conduct comprehensive analyses of the data often provide statistics that fall between these two extremes, though their numbers also differ considerably from one another and fall across a wide range. While the Bureau of Investigative Journalism puts the number as high as 26.5 percent, others estimate that the percentage of civilian casualties falls between 4 percent and 20 percent, and The New America Foundation put the number at a low of 8 percent.9
There is no way to settle these differences because often the drone strikes are in areas that are inaccessible to independent observers and the data includes reports by local officials and local media, neither of whom are reliable sources.10 The most cited statistics on the drone strikes in Pakistan-a data set compiled by the New America Foundation and Peter Bergen- relies completely on media reports.11 It is a problem that plagues a majority of the media stories on any particular strike: estimates of civilian casualties are often based upon other media reports, producing what the Human Rights Clinic at Columbia Law School calls "an echo chamber" effect.12
In short, there is no fully reliable-or even highly reliable-way to determine the ratio of civilian to militant casualties caused by drone strikes. …