Why you should be (very) skeptical of Amnesty International’s report

In a recent post, “Breaking the Silence on Civilian Casualties from U.S. Air Strikes in Somalia,” Amnesty International official Daphne Eviatar extols her organization’s new report which claims to have “documented 14 civilians killed, and eight injured, by just five U.S. strikes in the Lower Shabelle region of Somalia between April 2017 and December 2018.” According to Amnesty, the “attacks appear to have violated international humanitarian law, and some may amount to war crimes.” Allow me to explain why I believe you should be very skeptical of Amnesty’s charges.

Amnesty got it wrong…again

Amnesty’s report predictably led to headlines like this: “Human rights group accuses AFRICOM commanders of war crimes, calls for investigation,” but U.S. Africa Command (AFRICOM) “closely reviewed all information available and found no information to corroborate the allegations of civilian casualties highlighted in the report.” In her post, Eviatar lists four reasons why there is “discrepancy” between her organization’s casualty assessments numbers and those of AFRICOM. All of her reasons cast aspersions in some way on the people at AFRICOM or, as she imagines, “another US government agency — the CIA.”

Permit me to add a fifth possible reason for the discrepancy: Amnesty got it wrong…again.

In their new assertions, Amnesty cites (no less than five times) their 2013 report about drone strikes in Pakistan. However, that report was largely debunked by journalist David Axe in his essay, “Dear Amnesty International, Do You Even Know How Drones Work?” Axe notes that:

“It seems that some Pakistanis in the tribal areas describe all warplanes and even helicopters as “drones”—and as American. They apparently just assume that explosives raining from the sky necessarily come from robotic planes flown by the U.S.”

Axe is especially convincing when he points out that several of Amnesty’s so-called “eyewitness accounts” are actually “inconsistent with known drone tactics and the well-understood limitations of unmanned aircraft in general.” For example, Amnesty says that one of their “witnesses” claimed that the “drone planes were flying over our village all day and night, flying in pairs, sometimes three together.” (Emphasis added.)

Anyone familiar with drones would have to agree with Axe when he says that drones, like “Predators and Reapers necessarily operate alone, because their pilots—sitting in trailers thousands of miles away in Nevada, monitoring their ‘bots’ flights through narrow camera lenses—lack the visibility to fly in close formation with other aircraft.”

Additionally, consider this from Axe’s critique:

“People are mentally disturbed as a result of the drone flights,” the Esso Khel villager told Amnesty. “We can’t sleep because of the planes’ loud sound.”

Problem is, Predators and Reapers spend most of their time at altitudes where they are generally inaudible. It’s for that reason that the U.S. Army used Predators in Iraq to orbit over insurgents’ meetings, striking them only after all the fighters had gathered together.

In other words, six years after the fact Amnesty is still relying upon an obviously flawed report to support their new claims. Unlike the U.S. military which will admit when it gets something wrong and will re-open closed civilian casualty investigations when new information comes to light, it seems that Amnesty just can’t bring themselves to do so, even when the evidence is as crystal clear as it is in Axe’s analysis.

Eviatar also characterizes the casualty number “discrepancy” as being between her organization and AFRICOM, but that isn’t really accurate either. As of this writing (March 31st) the Bureau of Investigation Journalism (BIJ) – hardly a U.S. government mouthpiece – shows that during the period of 2017 to 2018 their “best estimate” is that between zero and 24 civilians were among the 447 to 577 people killed in 80 strikes. Obviously, the “zero” casualties AFRICOM estimates falls within BIJ’s “best estimate.”

Transparency about the “witnesses”?

Journalists should be wary of Amnesty’s claims given their almost complete lack of transparency about their main witnesses.   Amnesty removed from the report “information that could identify interviewees, or members of their communities” with whom they spoke. Where people were allegedly “quoted,” pseudonyms were used, and even the “specific dates and locations of interviews” were also withheld. In short, there is no way a journalist or anyone else can verify what was said, or whether the alleged “witnesses” even exist.

Amnesty says “[s]ecurity concerns and access restrictions prevented Amnesty International from conducting onsite investigations and severely limited the organization’s ability to freely gather testimonial and physical evidence.” (Emphasis added.) One would think that being “severely limited” in the ability to gather evidence would militate against even insinuating war crimes against anyone, but not so with Amnesty.

Amnesty also used unidentified “individuals living in Somalia” who they “hired” in order, Amnesty says, “to locate survivors and witnesses and gather information about specific air strikes.” In other words, the hirelings were paid not to dispassionately determine facts as to what may have happened (or not) in a particular circumstance but rather were specifically tasked to find “survivors” of what Amnesty pre-determined were “air strikes.”

In describing its methodology, Amnesty makes some curious statements. It says that “[n]o incentives or monetary compensation were provided to interviewees in exchange for their accounts,” but doesn’t say whether or not it was suggested to the interviewees that they might get something in the future.

Consider that Amnesty says that “interviewees were informed about the nature and purpose of the research, as well as how the information they provided would be used.” The information was, in fact, “used” in a report that some 14 times employs the word “compensation,” often in the context of a demand that the U.S. to “offer compensation and explanation to survivors and families of civilians killed in US strikes regardless of whether or not the casualties were caused by a lawful attack” – something not required of the U.S. by international or domestic law.

You be the judge: if the interviewees really were told their “accounts” would be “used” in a report demanding the U.S. “offer compensation,” do you think that would incentivize “interviewees” to be sure their statements would put them among the “survivors and families of civilians” who would get that compensation?

Furthermore, Amnesty said it made “encrypted voice calls placed from phones located outside Al-Shabaab-held territory” to persons whom their paid “individuals” said were “survivors.” Of course, there is no real way for Amnesty or anyone else to verify the identity of the person with whom they were speaking, or whether that person was under coercion by al-Shabaab.

Al-Shabaab lawfare

What is ironic about Amnesty’s dependence upon phone calls to people inside al-Shabaab controlled territory is that a UN report about al-Shabaab’s sophisticated propaganda campaign noted that the terrorists “give out free phones” to their members. In its response, AFRICOM notes the obvious:

“It is in the interest of the terrorist group al-Shabaab to untruthfully claim civilian casualties. It is also in the interest of al-Shabaab to coerce community members to make untrue claims.”

In essence, AFRICOM is saying that al-Shabaab is practicing what I call lawfare. Writing in a chapter entitled “Reframing Lawfare” in the 2017 Netherlands Annual Review of Military Studies Dutch colonel Joop Voetelink explains how this form of lawfare works:

[A party to a conflict will] attempt to create the image that their opponents “cause disproportionate collateral damage and civilian casualties, thereby undermining the perception of legitimacy”. Tactics based on this concept are designed to have a double effect on a law-sensitive opponent. Firstly, it may lead an opponent to impose additional constraints on his own actions whenever civilian casualties or damage to civilian property are to be expected, rendering its forces less effective. Secondly, the tactic may cause a loss of popular and international support for allegedly operating in violation of the laws of armed conflict.

Let me be clear with this warning as well: too many human rights organizations seem to mistakenly believe that if the U.S. and other rule-of-law countries can be constrained in – or barred from – the use of force, civilian lives will necessarily be saved. As I’ve explained before, it’s quite likely that undue constraints could produce the exact opposite result. How? Terrorists who might have been neutralized in a lawful airstrike will live to impose their horror on more of those unable to defend themselves than civilians who might have been lost had the strike taken place.  Let’s not forget that al-Sabaab “violently persecutes non-Muslims and clashes frequently with humanitarian and international aid workers” and metes out “amputations, and beheadings” as regular punishments.

Which is better, the U.S.’s or Amnesty’s “satellite imagery and data”?  

AFRICOM also says that its review is “based on post-strike analysis using intelligence methods not available to non-military organizations.”

However, in an apparent effort to bolster its credibility, Amnesty says it too uses “satellite imagery and data.” However, no serious analyst would say that commercially-available satellite imagery is as precise as that available to the U.S. military. Amnesty does not even claim to have the signals intelligence or other intelligence methodologies available to the military. The U.S. spends billions to maintain the world’s most advanced surveillance and reconnaissance capability. Does anyone think a non-governmental entity could match it?

“No, solely is not accurate”

Furthermore, to support their assertions Amnesty makes much of an interview with retired Army brigadier general Don Bolduc who served as the Commander of the Special Operations Command in Africa from April 2015 to June 2017. Specifically, Eviatar (and Amnesty) claim that BG Bolduc told them “AFRICOM considers individuals lawfully targetable based solely on a) age, b) gender, c) location, and 4) (sic) geographical proximity to Al-Shabaab.” The Amnesty report uses this to claim that U.S. forces “appear to be acting in violation of IHL.”

Amnesty concedes that AFRICOM stated that Bolduc’s “purported articulation of targeting standards does not accurately reflect the targeting standards of AFRICOM or [Department of Defense]’,” but dismisses that denial – evidently wanting readers to believe such factors “solely” determine targeting.

However, “solely” didn’t track with anything I knew about targeting based on my many years of military service, so I contacted Bolduc about his “purported articulation of targeting standards.”  Unsurprisingly, he told me via email that:

“No, solely is not accurate. As you know, there are other factors involved in the targeting approval process. So, solely, is not an accurate representation of what I said or meant.” (Emphasis added.)

Bolduc also said that AFRICOM commander, General Thomas D. Waldhauser, “provided guidance and was clear about protecting civilians.” There are, indeed, other factors involved in targeting, and a sampling of the complex collateral damage methodology (CDM) aimed at ensuring compliance with the law can be found here.

Additionally, Bolduc told me that he is “all about targeting al Shabaab, but [he did] not support this method,” as he simply believes that airstrikes alone will not defeat the terrorist group. AFRICOM seems to agree; the Military Times reports that ”[t]raining advanced infantry — not airstrikes — is AFRICOM’s primary effort in Somalia.”

In any event, Bolduc’s view as to the most effective strategy is something upon which reasonable people can differ, but it’s not the same as alleging that the attacks may amount to war crimes.

In fact, Bolduc explicitly told me in an email that he did not “accuse anyone of committing war crimes” adding that “as a matter of fact, [he] went out of [his] way to ensure [Amnesty] understood [that he] did not support [that] assertion.” Bolduc said he “honestly thought [Amnesty] wanted [his] views to clarify differences in the policies and better understand why this would open the aperture for increased targeting.” Again, “increased targeting” is not suggesting unlawful behavior.

The desire to clarify the differences in the policies fits with what Bolduc told the Daily Beast in a November 2018 interview. There, he was explaining that the Obama administration’s policy of requiring that there be a “near certainty that non-combatants will not be injured or killed” had been changed to ““reasonable certainty” that civilians would not be harmed. Bolduc does seem to disparage that change:

“If you can say we’re not gonna have any civilian casualties, then the senior leader, a senior general officer, can say ‘OK, take the shot’,” Bolduc said. “And he’s not going to be questioned by the Pentagon or anybody else if later in the assessment the target wasn’t there but we ended up killing 8-10 al Shabaab fighters and lower level leadership.”

But the key point is that the law of war doesn’t require even a “reasonable certainty” that civilians would not be harmed. Instead the law says that “incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof” must not be “excessive in relation to the concrete and direct military advantage anticipated.” (Emphasis added.)

The legal standard is “reasonableness,” not “certainty”

Do commanders need to be “certain” that the incidental civilian losses “anticipated” in an attack will not be “excessive”? No. Clausewitz tells us that “War is the realm of uncertainty; three quarters of the factors on which action in war is based are wrapped in a fog of greater or lesser uncertainty.”  Commanders often have to make decisions based on imperfect information, and the law accepts this reality.

Rather than certainty, the standard is, as Ian Henderson and Kate Reece explain, that of a “reasonable commander.” They observe that in the “Galić [case] the International Criminal Tribunal for the former Yugoslavia (ICTY) held that”:

“In determining whether an attack was proportionate it is necessary to examine whether a reasonably well-informed person in the circumstances of the actual perpetrator, making reasonable use of the information available to him or her, could have expected excessive civilian casualties to result from the attack.”

Put another way, even if a commander is not “certain” the casualties will not be excessive, there is no violation of the law of war so long as he or she reasonably believes they will not be excessive based on the reasonable use of the information available to him or her.  Thus, the policies of both the Obama and Trump administrations in this respect were discretionary, political decisions that demanded more than what the law of war requires.

All of this discussion illustrates the high evidentiary bar that must be reached before a prima facie war crime charge can be made out. For example, Article 8 of the Rome Statute of the International Criminal Court criminalizes:

Intentionally launching an attack in the knowledge that such attack will cause incidental loss of life or injury to civilians or damage to civilian objects or widespread, long-term and severe damage to the natural environment which would be clearly excessive in relation to the concrete and direct overall military advantage anticipated. (Emphasis added.)

The Amnesty report has exactly zero information as to what knowledge AFRICOM commanders did or did not have before the particular strikes they highlight, or even what the “overall military advantage” the commanders specifically “anticipated.” Without credible evidence of the required nefarious intent and knowledge, no war crime is established. Innuendo and speculation do not suffice. And, no, indications of between zero and 24 civilians in 80 strikes is hardly adequate circumstantial evidence of it; to the contrary, it indicates that feasible precautions were, in fact, taken.

Non-state actors as lawful targets

Additionally, Eviatar claims that “AFRICOM takes an overly-broad view of who qualifies as a combatant or lawful target under international law.” This is simply wrong. Although it appears that AFRICOM is, in fact, limiting its targets to those terrorists overtly engaged in combat operations, it is still important to clarify what the law of war would actually permit.

In this respect Amnesty seems to take the position that the only valid targets among non-state actors organized into an armed group engaged in continuous combat operations are those physically conducting attacks. That view has long been rejected by many countries, much because of its unfairness to members of the armed forces of nation-states, but also because it denies the realities of 21^st century combat. We should not be providing legal protection to those who play key roles in campaigns of terror like al-Shabaab’s.

The U.S. Department of Defense Law of War Manual sets forth an analytical framework (¶ 5.7.3) for determining when “individuals who are formally or functionally part of a non-State armed group that is engaged in hostilities may be made the object of attack because they likewise share in their group’s hostile intent.” (Emphasis added.)

That framework rightly points out that “performing a function for the group that is analogous to a function normally performed by a member of a State’s armed forces” is “information” that “might indicate that a person is [formally] a member of a non-State armed group.” Similarly, “performing tasks on behalf of the group similar to those provided in a combat, combat support, or combat service support role in the armed forces of a State” are indicators that “a person is functionally a member of a non-State armed group.”

The Manual variously cites scholar Ken Watkin’s logic. Watkin, a former Canadian Forces judge advocate, explains (p. 229):

“[Terrorists] need a logistical infrastructure, which consists of individuals … who provide everything from reconnaissance of the potential target … to the provision of a safe house and food, and the explosives-laden vehicle or suicide belt.’ … To limit direct participation to persons who place or detonate explosives is an artificial division of what is fundamentally a group activity. … The person who is key in planning and facilitating such deadly attacks must be a valid target as a direct participant in hostilities…” (Emphasis added.)

Watkin also says (p. 231):

“The argument that civilians are protected unless engaged in overtly aggressive acts like carrying weapons may be particularly difficult to maintain where armed groups are technically accorded civilian status by virtue of not being considered lawful combatants. To the extent that civilians fulfill the same function as combatants, either in the armed forces or as part of the organization of an ‘illegitimate’ nonstate actor, they are logically subject to targeting under the same provisions of international humanitarian law.” (Emphasis added.)

Despite Amnesty evidently thinking that combat support roles are not integral to al-Shabaab’s infliction of terror, a new (Mar 25) report concludes otherwise:

“In al-Shabaab’s campaign of terror, there are women who play key roles. They are not mere victims, but active players who the group heavily relies on for non-combative roles. These are indirect roles and are viewed as more important for the survival of the group than the female actors moving to the frontlines as combatants.” (Emphasis added.)

The same report points out that it is “now emerging that women are also at the center of one of Africa’s deadliest insurgencies Al-Shabaab.” Similarly, in 2017 Al Jazeera reported that “UN Secretary-General Antonio Guterres expressed alarm at reports that children may constitute a large part of the force recruited and used by the armed al-Shabab group in Somalia,” adding that the “UN chief said al-Shabab used children in combat, with nine-year-olds reportedly taught to use weapons and sent to frontlines.” In June 2018 the Borgen Project said that “Al-Shabab has recruited and trained children as young as age nine to be combatants” and that “over 50 percent of al-Shabab members are believed to be children.”

The tragic fact of child soldiers caused Canada to formalize its approach to the dilemma. As the Economist reported in 2017:

“On March 2nd Canada adopted a military doctrine that explicitly acknowledges soldiers’ right to use force to protect themselves, even when the threat comes from children. “A child soldier with a rifle or grenade launcher can present as much of a threat as an adult soldier carrying the same armament,” it says.”

The Economist further noted:

“The authors of the new directive seem to be aware that a policy to shoot child soldiers even in self-defence could provoke outrage. So far, human-rights groups have expressed understanding. Canada is trying to strike a balance between treating children as innocents and recognising them as battlefield threats, says Jo Becker, a children’s-rights specialist at Human Rights Watch in New York.”

Again, there is no evidence that AFRICOM is targeting child soldiers, but the logic of using force to protect those being victimized by al-Shabaab terror applies. The point here is that even if there were reports of women and children as young as nine being killed, that circumstance does not, ipso facto, mean the law of war was violated.  Moreover, Amnesty never even mentions Article 58 of Protocol I which obliges al-Shabaab to remove civilians from the vicinity of military objectives.

Amnesty’s confused understand of the law of targeting

Amnesty has a confused understanding of the law of targeting when it comes to determining who is in enemy armed group legally subject to attack. To illustrate, last year in a Just Security podcast on “Military Targeting under Trump” Eviatar made this claim:

“[I]f you look, for example, at the law of war manual from the Defense Department, it allows the government to assume that someone’s a member of an armed group who could be killed if they’re traveling with the armed group or if they stay at the guest house that the armed group frequent. So an extremely loose definition and that is putting a lot more people in the crosshairs of lethal force.”

I urge readers to look at the actual text of ¶ 5.7.3 “Persons Belonging to Non-State Armed Groups” of the Law of War Manual (extract found here) and you’ll find that it simply lists the “types of information might indicate that a person is a member of a non-State armed group” and includes “accessing facilities, such as safehouses, training camps, or bases used by the group that outsiders would not be permitted to access.” (Emphasis added.) 

The sole reference to “guest house” is found in footnote 212 which compares the listed criteria to the conclusions of Federal courts as to evidence sufficient to tie a detainee to membership in an enemy terrorist group. Here it is in its entirety:

Cf. Alsabri v. Obama, 684 F.3d 1298, 1306 (D.C. Cir. 2012) (“[I]t is difficult to believe that ‘Taliban fighters would allow an individual to infiltrate their posts near a battle zone unless that person was understood to be a part of the Taliban.’”) (quoting Alsabri v. Obama, 764 F.Supp.2d 60, 94 (D.D.C. 2011)); Uthman v. Obama, 637 F.3d 400, 406 (D.C. Cir. 2011) (“In two prior cases, this Court has stated that staying at an al Qaeda guesthouse is ‘powerful— indeed ‘overwhelming’—evidence’ that an individual is part of al Qaeda. Al–Adahi, 613 F.3d at 1108 (quoting Al– Bihani v. Obama, 590 F.3d 866, 873 footnote 2 (D.C. Cir. 2010)) (alterations omitted). The reason for that assessment is plain: It is highly unlikely that a visitor to Afghanistan would end up at an al Qaeda guesthouse by mistake, either by the guest or by the host.” (Emphasis added.)

Plainly, the facts are rather different from how Eviatar sought to characterize them.

The “grave danger” of an after-the-fact, “effects-based” standard

Amnesty’s attempt to use post hoc “body counts” to make instant assessments of criminality is the kind of approach that Professor Laurie Blank roundly criticizes in a chapter of a new book, Complex Battlespaces: The Law of Armed Conflict and the Dynamics of Modern Warfare. Blank makes it clear that the law finds war crime culpability not simply based on an after-the-fact count of casualties from an attack, but rather only when the evidence includes a finding that the commander departed from the standard of reasonableness and good faith in conducting it. Most importantly, she further warns about the grave damage that an after-the-fact, “effects-based” standard would have on the law of war:

“[T]he most direct and evident consequence of the effects-based approach is that it opens the door to a grave danger: the exploitation of the law by the defending party for its own defensive and propaganda purposes. If the results of an attack determine the lawfulness of that attack, the defending party’s precautionary obligations are emasculated because they no longer factor into the legal assessment of who bears responsibility for the harm to civilians.”

‘When parties face no legal consequences, and a potential operational advantage, for co-mingling civilian and military objects, every apartment will be a command center as militaries and armed groups embed themselves in cities to use the civilian population as a shield.’ But these tactical advantages are merely the beginning. These tactics often have a more problematic, strategic purpose: to use the resulting civilian deaths as a broader strategic tool to level accusations of war crimes, diminishing support for the war effort and the overall legitimacy of the military operation.”

“Legal assessments based on effects merely ratify the use of civilians and the civilian population as a shield for military operations and—albeit unintentionally, of course—directly undermine the very purpose of the LOAC’s core principles of distinction, proportionality, and precautions.” (Emphasis added; citations omitted)

Blank is, of course, exactly correct as it’s beyond cavil that al-Shabaab makes extensive use of human shields (see e.g., here and here). Reports like Amnesty’s could incentivize terrorists to burrow ever deeper into civilian population as they can see that doing so can get their hunters branded as “war criminals.”

How human rights groups can help – but also hurt – human rights

As the U.S. military has said on many occasions, non-governmental organizations can help to make combat operations more protective of civilians. This includes raising concerns about specific incidents and furnishing evidence for further review.  As I noted above, the U.S. will re-open closed civilian casualty investigations when new information comes to light.

Constructive criticism is useful, but Amnesty’s report is not that. Instead, for many reasons Amnesty’s report rightly earns suspicion and skepticism. What is really disturbing to me is what I see as a rather cavalier attitude about impugning the professionalism of those in uniform by raising the “war criminal” label. In my view, any human rights organization ought to be scrupulously adherent to the human right to the presumption of innocence, and refrain from suggesting criminality until the evidence is clear. Shouting “war criminal” may garner headlines helpful to NGOs in many ways, but some things are just wrong to do.

Still, as we like to say on Lawfire,^® gather the facts, assess the law and the arguments, and decide for yourself!