Afghanistan Analysts Network – English

International Engagement

Drone warfare 1: Afghanistan, birthplace of the armed drone

Kate Clark 12 min

Using drones to carry out targeted killings has become an integral part of the United States’ ‘war on terror’. Afghanistan in the late 1990s was the laboratory where the US developed armed drones as it searched for a way to deal with Osama bin Laden who was then ordering attacks on American targets from his safe haven in Kandahar. At that time, Washington was uneasy about ordering an assassination, especially one likely to result in civilian casualties. After 9/11, such doubts disappeared and it embraced drones, using them to carry out targeted killings of Islamist militants in many countries. In this first of two dispatches, AAN’s Kate Clark looks at armed drones in Afghanistan.

A Reaper drone comes into land at Kandahar Airbase (Flying Officer Owen Cheverton: 2009)A Reaper drone comes into land at Kandahar Airbase (Flying Officer Owen Cheverton: 2009)

A second dispatch will look at the expansion of America’s targeted killing by drone programme in the war on terror and asks whether Afghanistan might in the future see a US ‘drone-only’ or ‘drone-mainly’ mission of the sort seen in Pakistan’s tribal areas.

Bin Laden and the birth of drone warfare

The project to create armed drones grew out of the need felt by Washington to eliminate the threat posed by bin Laden. In the late 1990s, he was orchestrating attacks on American targets while living under Taleban protection. US options were limited by a presidential standing order banning assassinations (1) which meant the CIA was legally bound to plan an operation with detention as its sole aim. Additionally, CIA officials were worried about the women and children living in bin Laden’s compound – visible on satellite footage – who would be harmed if the capture operation turned into a firefight. “[CIA officers,” reported Steve Coll, “found themselves pulled into emotional debates about legal authorities and the potential for civilian casualties…” (2) (p 393)

In 1998, bin Laden ordered attacks on two American embassies in east Africa and Washington responded with Cruise missile strikes on training camps in Khost (which it said was an act of self-defence, not an assassination attempt). Even after that, however, Washington hesitated about making another attempt to kill the al Qaeda leader. The CIA insisted on definitive legal cover from the White House that officers would not later be charged with having carried out an ‘illegal’ assassination. Uneasiness was exacerbated by the fact that officials could never identify bin Laden with enough confidence to go ahead with a missile strike and they were still worried about killing women and children. Ahmad Shah Massud’s intelligence aides scorned this hesitation, reported Coll, portraying it as the US insisting on “capturing the king without disturbing the pawns.” (p535)

It was the need for accurate, absolutely up-to-date information about the target that drove the development of armed drones. They had already been used for surveillance in the Balkans, but now the decision was taken to increase their range and reliability and to arm them. Shortening the wait between target identification and strike, it was thought, would reduce the possibility of a precise strike with minimum ‘collateral damage’.

In the end, killing bin Laden by drone became feasible just as al Qaeda launched its attacks on New York and Washington on 11 September 2001. The US responded with a full military assault on al Qaeda and its hosts, the Taleban, and the first armed drone strike came that autumn, in Afghanistan, with an attempt to kill Taleban leader Mullah Muhammad Omar (see here). This time, there was no debate about legality: this was not now an assassination, but the lawful killing of a combatant during wartime.

Armed drones have been used ever since by the US in its ‘war on terror’, for both targeted killings and air support for troops on the ground. They have been deployed in Afghanistan, in Iraq after the 2003 invasion and more recently in Libya, and against the Islamic State in Iraq and Syria (for information about strikes in these three countries, see the Airwars website). Drones have also been used for targeted killings of al Qaeda and what the US calls ‘associated forces’ in Pakistan, Yemen and Somalia. Targeted killings by drone have become, former director of the CIA Michael Hayden said, “part of the American way of war.”

Since 2001, the US has asserted its legal right to kill hostile non-state actors if their host government is “unwilling or unable” to deal with the threat. The concerns which agonised the CIA and the White House over killing bin Laden, before 9/11 – whether targeted killings were legal and the danger of civilians being harmed in an assassination attempt – are now dismissed or downplayed by Washington. (This will be looked at in more detail in the second dispatch in this mini-series).

Drones in Afghanistan

Armed drones have been flown for more than fifteen years in Afghanistan. Yet data about them is scarce. Statistics for all aircraft flown by the United States, which, since 2014, has been the only foreign state carrying out combat operations in Afghanistan, are not disaggregated; we know only the numbers for the total air sorties flown and munitions dropped by all US aircraft. It had been believed that the statistics – even though not disaggregated to distinguish drones from other aircraft – were good. The US Air Force has been collating and publishing them for several years. (Note for anyone following this issue: the URL has recently been changed slightly; this one works.) However, the Military Times recently revealed that data for aircraft operated by the army in Afghanistan, Iraq and Syria has been excluded from these statistics. The discrepancy is large: in Afghanistan, in 2016, for example, 615 strikes had been reported, but including strikes from army aircraft, the number rose to 1,017. Still, although some of those extra strikes will have been by drones – the Military Times reported that the US army was flying MQ-1 Gray Eagles to help provide “lethal support” (3) – it seems the bulk of the extra airstrikes came from Apache helicopters.

Working out who flies what and under whose command is difficult. Generally, air strikes can be ordered, according to a military spokesman for US forces in Afghanistan who spoke to the Military Times, for “self defense, counter terror and strategic effects, which may be required when senior commanders believe U.S. firepower could help turn the tide in regions deemed vital to Afghanistan’s broader stability.” Senior commanders have told AAN this, for example, might be preventing a district or provincial centre falling to the Taleban. Those who have access to airpower, including drones, on the US side are: the US Air Force, the Joint Special Operations Command (JSOC), which is the key player in counter-terrorism and particularly US targeted killings operations in Afghanistan, and the army. The CIA has a close working relationship with JSOC; in the past (and possibly still?), this included pooling intelligence and drawing up lists of targets for kill/capture operations. (4) The CIA also flies drones across the border to carry out targeted killings in Pakistan’s tribal areas. The use of covert Agency drones was important because of Islamabad’s claims to be hostile to the US strikes. (There will be more detail on this in AAN’s second dispatch on armed drones.)

Josh Smith, former Kabul correspondent with the independent US military newspaper, Stars and Stripes, and now with Reuters, has reported that the US Air Force flies MQ-9 Reapers and MQ-1 Predators out of American bases in Kandahar and Jalalabad. He told AAN that pilots and operators on the ground guide the drones in and out for take off and landing, but unless the mission is very local, for example, base protection, once a drone is in flight, control is handed over to other operators, almost all in the United States. The main hub is Creech Air Force Base in Nevada, but there are also a small number of control stations in other locations. “Pilots at Cannon Air Force Base in New Mexico, for example,” he said “flew most of the Air Force Special Operations Command missions.” Air force drone operators told Smith that they fly drones and conventional aircraft “for a range of customers.” That could include JSOC, the CIA (see also this piece) and the army, although all also have their own drones and pilots.

Between 2008 and 2014, before NATO’s transition to the non-combat, ‘train, advise, assist’ Resolute Support mission, the UK also flew armed drones in Afghanistan, with strikes made in 16 of the 34 provinces. These were “only,” said the Ministry of Defence, “in support of coalition ground forces”(see here). The Afghan Air Force now also has drones for surveillance. Members of the Taleban’s media branch recently posted pictures of what they say are their drones in action, carrying out surveillance in Afghanistan.

Drones and civilian casualties

Trying to determine the impact of drone strikes in Afghanistan is difficult. The scarcity of drone specific data and the sheer amount of other weaponry around makes singling out the effects of drones per se very difficult. Any air strike could have been carried out by a drone or another aircraft – it can be difficult to tell from the ground. Moreover, targeted killings have been carried out not only through airstrikes, but also in night raids (although here there is an aim to capture or kill). Both tactics proved controversial, but it was civilian casualties and the invasion of people’s homes that were upsetting, rather than the use of drones.

UNAMA, in its tracking of civilian casualties, does not separate those caused by drones from other aircraft. Indeed, it would not be able to disaggregate the figures without the US providing the information, if it wanted to.

There has been discussion about whether drones are or should be better than other aircraft in reducing the risk of killing civilians when making a strike. Supporters of drones and even some detractors point out that, because drones can loiter in ways that planes cannot, they allow for ‘tactical patience’ and more accurate targeting. All things being equal, therefore, there is less likelihood for civilians to be harmed. However, one study with access to classified military data which was able to compare strikes from drones and aircraft in Afghanistan (mid-2010 to mid-2011) found that drones were then causing ten times more civilian casualties. Co-author Sarah Holewinski of the non-governmental organisation Center for Civilians in Conflict, told The Guardian the disparity was a result of fighter pilots getting more training in avoiding civilian casualties: “‘These findings show us that it’s not about the technology, it’s about how the technology is used,’ Holewinski said. ‘Drones aren’t magically better at avoiding civilians than fighter jets. When pilots flying jets were given clear directives and training on civilian protection, they were able to lower civilian casualty rates.’”

There are other ways in which civilian casualties may be increased or reduced by the use of drones, although, here we can only speak speak about air power in general because the data for drones and other aircraft is not disaggregated. Reduction in international air cover generally since 2014, for example, has resulted, indirectly, in an overall increase in the number of civilians killed and injured in the Afghan conflict (see UNAMA report here and AAN reporting here. This is because the Taleban, no longer fearing attack from the skies, have been able to mass in ways that would previously have been suicidal. It has become possible for them to launch ground engagements and threaten Afghan towns and cities, pushing up the total number of civilians killed and injured in the conflict to new records.

However, the data also shows that having a ‘pro-government’ aerial capability does not necessarily safeguard civilians. In 2016, air strikes killed and injured more civilians than in any year since at least 2009 when UNAMA started systematic documentation. UNAMA attributed about 40 per cent of the 2016 casualties caused by air strikes to the Afghan air force and about 40 per cent to the US air force (with the other 20 per cent not attributable to either party, but caused by one of them) (See the UNAMA annual report here and AAN’s analysis here). The situation is grave enough for UNAMA to have called for an immediate halt to strikes by armed aircraft in civilian-populated areas and for “clear tactical directives, rules of engagement and other procedures” to be adopted. In many of those intervening years, far more sorties were flown and strikes made, so what is pushing the numbers up?

On the US side, successive military commanders had made it more difficult for air strikes to be ordered, putting in place new sets of precautions and conditions. These led to successive drops in the number of civilians killed and injured in air strikes (for detail, see here and here). It seems the sharp rise in civilian casualties caused by US air operations in 2016 was probably due to poorer intelligence (fewer ground troops means the US is not as knowledgeable as it used to be) and/or failure to follow procedures. The latter was the case with the strike on the Médecins Sans Frontières hospital in Kunduz in October 2015 and the strike on Buz-e Kandahari village, again in Kunduz on 2-3 November 2016 (in each attack, about 30 civilians were killed and another 30 injured). In both cases, UNAMA and others questioned whether the US had breached the Laws of Armed Conflict.

On the Afghan side, it seems casualties are coming from failures to distinguish civilian from military targets, including recklessly firing at Taleban despite the harm bound to come to civilians (again, both potential breaches in the Laws of Armed Conflict). On 9 August 2016 in Nawa Barakzai district in Helmand, for example, UNAMA describes how Afghan air force helicopters tracked Taleban into residential compound and fired munitions, killing and injuring women and children.

So, in 2016, we saw sharp rises in civilians killed and injured by the most sophisticated air force in the world and one of the least – Afghan air strikes are often from machine guns mounted on helicopters. It was not the weaponry itself that was the crucial factor, but rather, mission aims and guidelines, training and intelligence. As a factor determining the numbers of civilian casualties, at least currently in Afghanistan, whether armed drones are used or not, seems to be a minor consideration. Moreover, if they are causing any variation, the distinction is not visible to anyone outside the US military.

Living under drones

 Just one study has managed to delve at all into the impact of drones on communities in Afghanistan. In autumn 2015, field interviews were carried out for Durham University with people living in two (un-named) districts of Nangarhar which had seen a “surge” in drone use. It is a preliminary study (the authors are currently finishing a longer research paper), but nevertheless, gives some pointers to how Afghan civilians may experience drones. In the two districts, said the study’s authors, the population is largely Pashtun and has a “notable history of support and loyalties towards Kabul.” They found that:

Drone strikes receive widespread support, as long as they effectively and accurately target Taliban, Daesh and elements from Pakistan believed to create and perpetuate these groups. Indeed, these insurgent and terror-group elements are clearly seen by citizens in the fieldwork areas as their enemies – and enemies of Afghanistan.

Respondents believed drones were “supremely accurate” and praised them for instilling fear in Taleban and Daesh members, thereby disrupting their movement and activities. However, it was the simultaneous menace of both insurgent groups and drones on respondents’ everyday life which the authors of the study found revelatory:

Fear of becoming caught up in a drone strike as a result of running livestock, collecting firewood in the mountains or cultivating fields where insurgent groups hide and pass through is economically damaging and encouraging depopulation. Activities that manifest and reinforce important social ties and networks are curtailed by the presence and fear of drones. This includes: providing hospitality to strangers who visit homes and may turn out to be Taliban; gathering to celebrate weddings; observing funerals; discussing the day’s issues at night after subsistence work; and simply moving around the village after dark.

 In other words, even if drone strikes were successfully targeting Taleban and other combatants, and this was popular among civilians, this was not creating the conditions for anything like a resumption of ‘normal life’. Nor did the killings feel like a victory for those on the ground.

The character of drone use in places like Nangarhar, said the Durham University study “is becoming more like that across the border in FATA.” This refers to a model where drones are used for targeted killings in a narrow, counter-terrorism mission aimed not at stabilising Afghanistan, with an eye to “restoring and reinforcing viable and effective governance, social and economic structures,” but at “containing the ability of Afghanistan-based terror-related groups to commit acts of violence beyond its borders, especially in areas central to US and wider western interests.” These two aims have, since 2001, always co-existed in the international military mission in Afghanistan, but the authors of the Durham study believe that in places like Nangarhar the narrow counter-terrorism mission has become dominant.

Nangarhar is a particular case in Afghanistan, a province where the situation on the ground most closely resembles the tribal areas of Pakistan, with a widespread, but fragmented armed opposition, a relatively weak Taleban and a plethora of Afghan and foreign jihadist groups (see AAN reporting here. US military operations in Nangarhar are not particularly representative of how it operates elsewhere in Afghanistan. However, it may suggest what a narrow US counter-terrorism mission could look like.

Constraints on US combat operations were written into the US-Afghanistan Status of Forces Agreement (SOFA) (see here), signed in September 2014; they were limited to strikes on “al Qaeda and its affiliates.” This led to a concentration of US air strikes on Afghanistan’s eastern border with Pakistan, especially Nangarhar, where foreign jihadists are most active. (The US was also still playing its part in the non-combat NATO mission to support Afghan National Security Forces.) In June 2016, President Barack Obama, with Afghan agreement, did broaden out US targeting, allowing for strikes on Taleban and in support of Afghan troops and air strikes have been used more widely since. Even so, there are worrying aspects to US air strikes in Nangarhar. One to watch in 2017 will be civilian casualties. In 2016 UNAMA noted, there were “considerable increases in civilian casualties caused solely by international military forces in Nangarhar province.” 89 civilians were killed or injured in 13 aerial operations in 2016 compared to 18 during 10 aerial operations in 2015.

Looking ahead

For the most part, the use of US drones in the Afghan conflict has not been visible or controversial. This is in sharp contrast to their deployment in FATA, and also in Yemen and Somalia. In those three countries, where the US has conducted a ‘drone only’ or drone mainly’ mission to carry out targeted killings, there has been a fierce debate about their legality, effectiveness and impact on civilians.

A second dispatch in this mini-series will look at why and how the US drone programme expanded after that first drone strike in Afghanistan in October 2001. It will also examine the experiences of Pakistan, Yemen and Somalia and ask whether the ‘FATA model’ of drone use could be seen in the future in Afghanistan.

Edited by Sari Kouvo and Borhan Osman



(1) President Gerald Ford banned assassinations after various CIA scandals emerged after Watergate, including the Agency’s repeated attempts to kill Cuban president Fidel Castro. The ban was reinforced by Ford’s two successors. President Ronald Reagan’s version) said: “No person employed by or acting on behalf of the United States Government shall engage in, or conspire to engage in, assassination.”

(2) Steve Coll Ghost Wars: the Secret history of the CIA, Afghanistan and Bin Laden, from the Soviet Withdrawal to September 10, 2001, Penguin Books: London (2004).

(3) The military spokesman also included RQ-7 Shadows as one of the aircraft used by the army to provide “lethal support.” This model is actually used for surveillance, not targeting.

(4) In 2012, at the height of US kill or capture operations, the author wrote:

At the forward headquarters of the Joint Special Operations Command or JSOC at Bagram, where intelligence from multiple sources – the military, CIA, detainee interrogations, drone footage and intercepts – is collated, a joint targeting working group meets weekly. It has direct input from the Combined Forces Command and its divisional HQ, as well as lawyers, operational command and intelligence units, including the CIA and it places men deemed to be ‘insurgent leaders’ on the Joint Prioritised Effects List (JPEL) for capture or targeted killing.



CIA Drones USA