Remote Control: Remotely Piloted Air Systems

Written evidence from the Bureau of Investigative Journalism

Executive summary

ñ Britain is a key participant in drone warfare and is the only country the US has so far allowed to buy its armed Reaper drones. British pilots have conducted drone strikes using both British and US-owned drones in Iraq, Libya and Afghanistan.

ñ But little is understood publicly about the potential for civilian harm in drone strikes. This lack of transparency makes it hard for policymakers and the public to gauge their effectiveness.

ñ The UK has acknowledged only one drone strike in Afghanistan that has harmed civilians, and the Ministry of Defence says it does not collate comprehensive figures on civilian and insurgent deaths in the theatre. The US has also declined to publish overall casualty figures for its armed drone operations.

ñ However a recent analysis of classified military data relating to both US and UK aerial operations found that drones were "an order of magnitude" more likely to cause civilian harm than conventional air strikes.

ñ In light of concerning findings such as these it is doubly important that the British government establishes the international precedent of publishing a fuller record of drone strikes and their impact, to the extent that is operationally secure.

About the Bureau of Investigative Journalism
The Bureau of Investigative Journalism is an independent not-for-profit organisation that was established in April 2010. We pursue journalism that is of public benefit, undertaking in-depth research into the governance of public, private and third sector organisations and their influence.

Our work is philanthropically funded; the bulk of our funding comes from David and Elaine Potter. Further funding for specific drone projects comes from the Joseph Rowntree Charitable Trust and the Freedom of the Press Foundation, through a crowdfunding initiative. Although many journalistic organisations in the US use the philanthropic model, the Bureau of Investigative Journalism is the only unit of its kind in the UK.

Introduction

1. This submission will focus primarily on the current utility and dispersal of armed drones in both covert and military contexts. In particular, it will examine the lack of transparency surrounding the current usage of drones, and especially the lack of data concerning casualties.

2. Drones have been used in military contexts by the US and UK in Afghanistan, Iraq and Libya – situations of recognised armed conflict under the Geneva Conventions and International Humanitarian Law. The US has also conducted drone strikes in Pakistan, Yemen and Somalia – countries in which the legal basis for armed intervention is contested. These 'covert' strikes are conducted by the Central Intelligence Agency (CIA) and Joint Special Operations Command, a military unit whose operations are highly classified.

3. Senior US officials have described drones as highly precise weapons that are capable of targeting enemies including al Qaeda and affiliated organisations while causing minimal collateral damage. But independent organisations, including The Bureau of Investigative Journalism [1] (TBIJ), the New America Foundation [2] (a Washington-based think-tank) and the Long War Journal [3] (a US-based security studies blog), have charted hundreds of civilian casualties in covert drone strikes in Pakistan and Yemen. [4]

4. As a result of the work of TBIJ and others there is more detail in the public realm about supposedly covert drone strikes – such as those conducted by the US Central Intelligence Agency (CIA) in Pakistan – than about those conducted as part of military operations over the border in Afghanistan. This submission will discuss the findings of TBIJ’s two-year investigation into covert drone strikes.

5. It will also examine the key role played by British drones in the Nato-led International Security Assistance Force (ISAF) in Afghanistan. Despite having a comparatively small unmanned fleet in this theatre, British drones launch a high proportion of drone strikes in the conflict. But there is almost no information about the impact of these drone strikes, particularly on civilians. This lack of transparency means that policymakers and the public are deprived of the ability to gauge the effectiveness and accuracy of a weapons system that is widely expected to play a crucial role in future warfare. [5]

The US's covert use of drones|

6. Since 2011, TBIJ has tracked and investigated US covert drone strikes, using sources including credible media reports, legal affidavits, field investigations, research by NGOs, leaked Pakistani government documents and social media. According to TBIJ data the US has carried out over 400 such strikes in Pakistan, Yemen and Somalia since 2002. [6]

7. The US government has repeatedly described these operations as highly accurate, targeted attacks on militant groups that pose a threat to US interests, while also causing minimal harm to civilians. John Brennan, President Obama's former chief counterterrorism adviser and now director of the CIA, has praised drones' "surgical precision" and "unprecedented ability" to distinguish between terrorists and civilians. [7] President Obama's first public comments on the drone campaign included the assertion: "I want people to understand actually drones have not caused a huge number of civilian casualties, for the most part they have been very precise precision strikes against al Qaeda and their affiliates." [8]

8. Such drone strikes have reportedly killed senior al Qaeda leaders, as well as senior commanders of several other militant organisations including the Pakistan Taliban (TTP) and the Islamic Movement of Uzbekistan (IMU). They have also killed many hundreds of others who are reported to be members of militant organisations.

9. The legal basis for covert strikes has been questioned by international experts including Christof Heyns, UN special rapporteur on extrajudicial killings, who told a UN Human Rights Council conference in 2012 in a speech dealing extensively with US covert drone strikes: "Many targeted killings take place in circumstances that are both far from any recognised area of armed conflict, where the legal threshold[s] of 'armed conflict' are not met... Current targeted killing practices weaken the rule of law... they also set dangerous precedents for the future." He added: "Some states appear to want to invent new law, or stretch existing law beyond long-accepted understanding, in an attempt to justify extraordinary and often unlawful practices that are carried out in an attempt to meet short term goals." [9]

10. Senior officials have also questioned the legal justifications presented for covert drone strikes, including former CIA director Michael Hayden. "Right now, there isn't a government on the planet that agrees with our legal rationale for these operations, except for Afghanistan and maybe Israel," he told a journalist last year. [10]

11. The classified nature of covert drone operations means the lack of transparency surrounding them is perhaps unsurprising. Senior Obama administration officials and even the president himself have acknowledged the use of drones in Pakistan, Yemen and Somalia . [11] But there is no official recognition of individual strikes, even years after they have taken place. The US has only rarely provided overall estimates of drone casualties, and where it has done so it has often been in stark contrast to independent estimates. [12] This means that there is little detail against which to measure the US administration's claims that drones are an exceptionally accurate weapon that eliminate terrorist threats while causing almost no civilian casualties.

12. TBIJ’s research has included a particular focus on civilian casualties. The project was started partly because we felt the Obama administration’s repeated claims about the minimal collateral harm being caused by covert drone strikes were not being properly scrutinised. Defining who is a civilian in the regions where drone strikes take place is complicated by the fact that the drones target non-uniformed militant groups in parts of the world where adult men frequently carry weapons.

13. A New York Times report citing three dozen Obama administration insiders speaking on condition of anonymity claimed in May 2012 that the US had adopted a definition that classed "all military-aged males in a strike zone as combatants" unless they were later proven to be civilians. [13] TBIJ classes as civilians individuals who are not believed to be members of militant groups. Where the dead are described by researchers and reporters as "locals" or "tribesmen" (as opposed to "fighters", "militants" or "Taliban", for example) our methodology is to class these as potential civilian casualties. [14]

Pakistan

14. TBIJ has amassed reports of over 370 drone strikes in Pakistan since the first recorded attack in the country in 2004. These attacks have killed between 2,500 and 3,500, of whom between 407 and 926 are reportedly civilians and at least 168 are children. [15]

15. Working with Pakistan-based researchers, TBIJ has also carried out two field investigations that have corroborated multiple reports of drones returning to the site of earlier strikes to attack those carrying out rescue work. [16] In addition to killing alleged militants, several of these ‘double-tap’ strikes killed civilian. Following TBIJ's research the tactic was labelled a possible "war crime" by UN special rapporteur on extrajudicial killings Christof Heyns – an opinion that UN special rapporteur on counter-terrorism Ben Emmerson QC later said he "would endorse". [17]

16. The US government has consistently claimed that civilian casualties in drone strikes in Pakistan are very low. Early this year Dianne Feinstein, chair of the Senate Select Committee on Intelligence - one of the congressional bodies charged with overseeing the CIA’s use of drones - told a hearing that drone casualties from CIA strikes were ‘typically in the single digits’ each year. [18]

17. TBIJ’s findings contradict this claim. So too do the findings of other independent monitoring organisations including the New America Foundation and the Long War Journal, both of which also track numbers of drone casualties. According to the tallies of all three organisations, minimum civilian casualties have exceeded Feinstein's estimate for the four most intense years of the campaign – 2008, 2009, 2010 and 2011. Only in 2012 did a steep fall in reported collateral harm mean that reported civilian casualties were in single digits as Feinstein described.

18. The US administration’s estimate is also contradicted by the Pakistan government. In March this year Pakistan’s civilian government told UN special rapporteur on counterterrorism Ben Emmerson QC that at least 400 non-combatants - and possibly 600 - had died in drone strikes. This figure closely matches TBIJ's own estimate of confirmed civilian deaths. [19]

19. The CIA’s rate of drone usage in Pakistan has fallen sharply in the past 18 months. From a peak of 122 drone strikes in 2010, TBIJ has recorded fewer than 20 in January-August 2013. TBIJ's analysis finds that fewer people are now killed in each drone strike - in the first half of 2013 the average strike killed four people, a third of the rate of the same period of 2010. [20] The rate of civilian deaths has also declined, from a minimum of 100 in 2009 to no confirmed civilian casualties at all in 2013 at the time of writing. This coincides with increased public and political interest in drone strikes and particularly in civilian harm, as a result of which targeting policies appear to have been tightened.

Yemen

20. In Yemen, corroborating reports of drone strikes is more complex than in Pakistan as both the CIA and US military operate drones alongside US military fixed-wing (ie manned) aircraft, Yemen’s own air force and according to some reports also the Saudi air force. TBIJ has found reporting clearly indicating at least 54-64 drone strikes, alongside 82-101 other air strikes where it is unclear whether they were carried out by fixed-wing or unmanned aircraft. [21] At least 268 people - including over 21 civilians - have died in confirmed drone strikes, with a further 289 or more including 23 non-combatants dying in possible additional drone strikes.

21. The multiple overlapping aerial forces operating in Yemen means that it is difficult to conclusively establish accountability for individual incidents. At times it is clear that Yemen’s air force could not have carried out particular attacks because of the limitations of its ageing air fleet - a fact acknowledged by President Abdurahman Hadi, who told an audience in Washington last year his air force could not carry out night-time missions. [22]

22. The US government has at times denied responsibility for attacks that have later been shown to be the work of its armed forces. On occasion, this has meant that compensation to victims has been paid by the Yemeni government even though the strikes have later been shown to be carried out by the US. As a result families of victims have received far less compensation than the families of civilians killed in US operations in Afghanistan. [23]

23. What is clear is that the number of drone strikes has escalated in Yemen. In 2012 for the first time since 2002 the country saw more covert strikes than Pakistan (although many of these could not be confirmed as drone strikes), a trend that has continued into 2013. In July and August 2013, during the international terror alert that led many governments to close their embassies in the Middle East, there were nine confirmed drone strikes in the country in two weeks, killing 21-49 of whom at least six were civilians, including two children. While drone use appears to be declining in Pakistan, the increased activity in Yemen indicates there is little prospect of the non-battlefield use of drones ceasing.

Somalia

24. The continuing instability and violence that plague much of Somalia severely limits the reliable reporting of attacks in the country. Again tracking US covert drone strikes in the country is complicated by the presence of other forces: the US has carried out manned air strikes and cruise missile strikes, and the African Union Mission in Somalia (Amisom) operates a peacekeeping force comprising soldiers from a number of African nations. [24]

25. What news does emerge is frequently vague and it is possible that some incidents go unreported, both by the media and by bodies such as the UN’s monitoring group. [25] TBIJ has identified 3-9 US drone strikes in the country, killing between seven and 27 people since 2007. [26] No civilians have been confirmed killed in any of these incidents. However given the considerable challenges of reporting from the country it is possible that this record is far from complete.

26. A Bureau investigation has revealed that two individuals who died in US drone strikes in Somalia had previously been British citizens, but had lost their UK nationality under orders from the Home Secretary. [27] Mohamed Sakr, who was born in London, and his childhood friend Bilal al-Berjawi, who came to the UK as a baby, were stripped of their citizenship on national security grounds and were suspected of involvement in militant groups in Somalia. They died in two drone strikes a month apart in early 2012.

Military drones

27. Covert drone attacks represent a small proportion of overall drone strikes. Military data, where it has been published, is sometimes complex, reporting variously the number of strikes, the number of missiles fired (several missiles can be fired in a single strike) and the number of overall missions flown (most drone missions are flown for surveillance and do not lead to strikes).

28. According to official figures provided to TBIJ by the US Air Force, between 2008 and October 31 2012, together the US and UK carried out almost 1,200 drone strikes in Iraq, Afghanistan and Libya. [28] The same figures showed that the vast majority of military drone missions did not involve firing missiles: between January and October 2012, one in 30 sorties in Afghanistan led to a strike.

29. In October 2012, following discussions with TBIJ, the US Air Force started publishing data on how many missiles were fired by drones in Afghanistan, alongside similar statistics for overall air operations. However months later it reclassified the material on the grounds that they ‘disproportionately focused’ attention on drone operations. [29]

30. The statistics showed the increasingly important role played by drones in ISAF air operations: in 2012, almost one in eight missiles was fired by a drone, up from one in 20 the previous year. In January 2013, the last month for which the data was unclassified, 23% of all missiles launched in ISAF air strikes were fired by drones. [30]

British military drones

31. The UK is the only nation the US has so far allowed to buy and pilot its MQ-9 ‘Reaper’ drone. British drones started flying combat missions in Afghanistan in late 2008, piloting them from Creech Air Base in Nevada. In addition April 2012 the UK opened its first drone base on its own soil, at Waddington Air Base in Lancashire, from which British drone pilots fly combat missions in Afghanistan. [31]

32. In October 2012 the Royal Air Force (RAF) announced it was to double the size of its fleet from five to 10 Reaper drones within six weeks. [32] For comparison, although the US has declined to say how many combat drones it flies in Afghanistan, TBIJ estimates it is over 150.

33. In addition to flying RAF drones, British drone pilots regularly ‘embed’ with the US Air Force to fly US drones. [33] A Ministry of Defence (MoD) spokesman told TBIJ these embeds can last up to three years, with ‘fewer than 10’ pilots embedded at any one time. [34] Even when in a US embed, RAF drone pilots follow British rules of engagement, the MoD told TBIJ.

34. In April 2013 defence minister Andrew Robathan MP said in response to a parliamentary question that British pilots had been embedding for drone missions since 2006 and had flown US drones in ‘approximately 2,150 operational missions’ in Afghanistan, Iraq and Libya. [35] This was the first time it had been revealed that British pilots had flown drones in Iraq.

35. Figures released by Robathan in September 2013 show that British pilots launched 299 drone strikes between 2008 and July 31 2013. [36] The MoD later confirmed to the Bureau that this represented strikes launched from UK drones only, and did not include UK pilots operating in US embeds. An analysis of these figures against those provided to TBIJ by the US Air Force would suggest that Britain carried out 22% of all drone strikes in the theatre between 2008 and 2011. [37] The comparison shows the proportion of strikes carried out by British pilots increased steadily year on year. In 2011 - the last year for which comparison is possible - they launched 30% of all strikes.

Civilian casualties in British military drone strikes

36. As with covert drones, there is no official public accounting of either militant or civilian casualties. The RAF publishes details of selected operations in weekly bulletins: these sometimes include brief descriptions of drone strikes but do not always reveal how many people they were believed to have killed, often simply referring to ‘insurgents’ dying. [38] The bulletins are a ‘snapshot of current operations’, the Ministry of Defence (MoD) has confirmed, rather than a comprehensive record.

37. The United Nations Assistance Mission in Afghanistan (UNAMA) publishes annual and half-yearly reports on civilian casualties caused by both ISAF forces and insurgents. Since 2012, the first year for which data for drone strikes was published in a discrete category, it has recorded 12 verified strikes that killed 31 civilians; seven of these incidents were in the first half of 2013. [39] However when charting ISAF incidents it does not specify which force was responsible for any incident. In September 2013 the MoD informed the campaign group Drone Wars UK in response to a Freedom of Information request: 'we can confirm that we hold no information that suggests that any of the strikes listed in the UNAMA report [covering the first half of 2013] were carried out by British Reaper Remotely Piloted Air Systems (RPAS)'.

38. The MoD has made only limited disclosures about civilian casualties. In response to a Freedom of Information request made by Drone Wars UK in July 2011, it said there had been ‘one incident involving UK Reaper where there were six civilian casualties’ since July 1 2008. Also in July 2011, the MoD told the Guardian that four civilians had been killed in a March 2011 drone strike on a pick-up truck in Helmand province. [40] When contacted by TBIJ to explain the apparent discrepancy between the accounts, the MoD explained that there had been a clerical error in the response to the Freedom of Information request. It has not announced any further civilian casualties.

39. This appears to be an exceptional rate of accuracy. At the time of the July 2011 disclosures, British pilots had carried out well over 100 drone strikes – this would mean fewer than 1% of strikes had caused civilian casualties. For comparison, in Pakistan between 2008 and 2013, 18% of drone strikes have reportedly caused civilian casualties, according to TBIJ’s data.

40. A press officer told TBIJ in July 2013 that the UK does not collate figures on civilian casualties: ‘because of the immense difficulty and risks of collecting robust data’. It also declined to provide total casualty estimates .

41. However in July 2013, Dr Larry Lewis, principal research scientist at the US government-funded research organisation the Center for Naval Analyses and Sarah Holewinski, director of Civic, an advocacy group for victims of conflict, published an article in security studies journal PRISM that raised questions about how successful Coalition drones were in avoiding civilian casualties. The article was based in part on Dr Lewis’s analysis of classified military data on behalf of Coalition forces. Most of Lewis's findings remain classified: the study was carried out on behalf of the Joint Center for Operational Analysis, a military research unit. However the executive summary of the research asserted that drone strikes killed as many civilians per incident as fixed-wing strikes, and are ‘an order of magnitude more likely’ to kill civilians than fixed-wing aircraft - a startling claim that runs contrary to the depiction of drones as an exceptionally accurate weapons platform. [41]

42. Lewis claimed in an interview that his study had found that collateral harm was greater in drone strikes by a factor of 10. [42] This claim was initially reported in the press as saying drones killed 10 times more civilians than fixed-wing air strikes. TBIJ however understands that the correct interpretation is that drone strikes are 10 times more likely than fixed-wing attacks to kill civilians.

43. No public explanation has been offered as to why, according to this analysis, drones are so much more likely to cause non-combatant harm than fixed-wing air operations. Other analysts have questioned this finding, telling TBIJ it does not appear to tally with their understanding. Both the data analysed by Dr Lewis and the bulk of his findings remain classified, so it is not possible to scrutinise them.

44. The study was widely reported to pertain to US operations. However the data analysed in the study was for Isaf operations - meaning it included both British and US drone missions. Dr Lewis analysed a year’s worth of data, from mid-2010 to mid-2011. This coincides with a point when British drones were exceptionally active - as the data obtained by TBIJ shows, British pilots were responsible for between a quarter and 40% of all missiles fired by drones.

45. It is not known whether British-piloted drones are significantly more or less likely to harm civilians than US-piloted ones. The MoD claims it does not collate such data. However Dr Lewis's study indicates that data is being collated about civilian casualties in ISAF aerial operations. If British operations are not being recorded discretely this raises concerns over accountability for incidents of civilian harm. It also raises questions of whether British drone strikes – which made up a significant portion of the data analysed – are significantly more likely to cause civilian deaths than British fixed-wing air strikes.

Recommendations to the Committee

46. It is vital that Parliament and the public are able to evaluate whether British drone operations are as accurate as is claimed, and to assess the merits of this emerging weapons system. TBIJ therefore urges the Committee to encourage the Government to increase transparency around drone operations in Afghanistan, to the extent that is operationally secure. This is important not only in order to understand current operations, but also to set an international precedent for increased transparency around the use of drones.

47. Although the MoD claims it does not collate casualty data, it appears that at least some data does exist - that analysed by Dr Lewis. In the interests of transparency the MoD should publish the British portion of the data analysed by Dr Lewis, to the extent that is operationally secure.

48. In particular, the Committee should call on the MoD to clarify whether British-piloted drone operations are significantly more likely to cause civilian harm than British-piloted fixed-wing missions, and whether Dr Lewis’s findings are reflective of British drone usage.

September 2013