The Expert Panel’s Review of Amnesty International’s Allegations of Ukrainian IHL Violations

Last week, the New York Times reported that Amnesty International (AI) “sat for months on a report critical of the group after it accused Ukrainian forces of illegally endangering civilians while fighting Russia.” The report in question was prepared by a distinguished Panel of Experts that AI appointed to conduct a legal review of an AI August 2022 press release that had sparked widespread controversy. In an earlier Articles of War post, I assessed the allegations leveled by AI in its press release. Here, I offer my thoughts on the legal review of it by the Panel of Experts.

The Controversial Press Release

Amnesty International’s August 2022 press release accused Ukrainian forces of putting “civilians in harm’s way by establishing bases and operating weapons systems in populated residential areas, including in schools and hospitals, as they repelled the Russian invasion.” According to Agnès Callamard, AI’s Secretary General, the organization “documented a pattern of Ukrainian forces putting civilians at risk and violating the laws of war when they operate in populated areas.”

In response, AI was widely criticized, not least by Ukrainian President Zelensky, who charged the press release “shift(s) the responsibility from the aggressor to the victim.” His Foreign Minister similarly asserted that “This behavior of Amnesty International is not about finding and reporting the truth to the world, it is about creating a false balance between the criminal and the victim.” The press release even caused the organization’s Ukraine Director to resign in protest. Three days after its release, AI issued a statement announcing that it “deeply regrets the distress and anger that our press release … has caused.”

In my Articles of War assessment of AI’s press release, I concluded that some of the criticism was unfair, such as claims that AI should not be criticizing the behavior of a victim of Russia’s clear aggression. Such criticism ignores the principle of IHL’s equal application, which lies at the heart of that body of law. But I also felt that the organization went too far.

Amnesty International’s allegation of unlawful conduct by Ukraine is unconvincing. IHL is a nuanced body of law because it must carefully balance two sometimes competing interests – military necessity and humanitarian considerations. Even if the facts in the report are accurate, and there is no reason to believe they are not, it lacks the detail necessary to support the charge.

I labeled AI’s assertions “irresponsible.” Behind the scenes, this view was widely echoed in the IHL community. And Russia skillfully leveraged the allegations for lawfare purposes. For instance, soon after its release, Russia’s ambassador to the United Nations claimed, “We don’t use the tactics Ukrainian armed forces are using – using the civilian objects as military cover – (…) what Amnesty International recently proved in a report.”

Legal Review by an Independent Panel

To its credit, the organization appointed a Panel of Experts to provide a legal review of the press release as part of a broader evaluation into “what exactly went wrong and why.” Unfortunately, the Panel’s unanimous final report, which considered AI’s comments on a draft version, was completed on 2 February but not released publicly. As noted above, this led the New York Times to conclude that the AI Board had essentially buried it (see also The Guardian report). According to a source, the board decided to “merely use [the report] as one of several sources for a lessons-learned document to circulate internally” rather than releasing the whole document, or at least excerpts. Whatever the reason, the failure to make the report public is highly disappointing, especially for an organization that often criticizes States for their lack of transparency.

The report was authored by five distinguished scholars: Emanuela-Chiara Gillard (University of Oxford), Kevin Jon Heller (University of Copenhagen), Eric Talbot Jensen (Brigham Young University), Marko Milanovic (University of Reading), and Marco Sassòli (University of Geneva). I know them all well on both the professional and personal levels. It is a star-studded group with truly exceptional IHL expertise and unquestionable character.

My only hesitation regarding composition when the group was formed (and it was but a slight one) was that it would benefit from more operational expertise, which can be helpful in applying IHL’s highly contextual rules. Indeed, with respect to the delicate balance between military necessity and humanitarian considerations that infuses IHL, the group might be said to lean slightly in the direction of the latter.

But the proof is in the pudding. The report is a model of objectivity. As explained below, the Panel defended AI when merited and criticized the organization where it fell short. And their assessments are operationally sound.

The Legal Findings

In preparing its report, the Panel had substantial access to internal AI deliberations about the press release, including confidential evidence that informed it, and conducted extensive interviews with AI personnel. AI asked the Panel to address two questions based on this material.

1. Did the legal analysis in the Press Release fall within an acceptable range of interpretation of IHL?

2. Was the evidence underlying the Press Release sufficient to support the legal conclusions that AI reached?

The Jus ad Bellum and IHL

The Panel began by addressing the relationship between the prohibition on the threat or use of force found in Article 2(4) of the UN Charter and IHL. Pro-Ukrainian sentiment, which is well-justified, led some critics of the press release to disparage it on the basis that Ukraine was merely defending itself pursuant to Charter Article 51. That was not an appropriate criticism. As the Panel emphasized, the principle of equal application means IHL applies to all parties to a conflict; it is irrelevant whether they are using force lawfully or unlawfully.

This principle was most famously recognized in the 1948 Hostages (U.S. v. List et al.) case before a post-war U.S. Military Tribunal, which held,

Whatever may be the cause of a war that has broken out, and whether or not the cause be a so-called just cause, the same rules of international law are valid as to what must not be done, may be done, and must be done by the belligerents themselves in making war against each other, and as between the belligerents and neutral states.

It also appears in the Preamble to Additional Protocol I, to which Ukraine and Russia are Party.

Based on the principle, the Panel found that the outcry against criticism of Ukraine for violating IHL on the basis that it is the victim is “legally and morally unjustified” (¶ 10). However, it suggested that AI should have discussed the principle of equal application and “for public perception reasons, the more serious and widespread character of the violations of IHL by Russia should have been mentioned” (¶ 16). In retrospect, that is sound advice, although, in fairness, AI had earlier, and frequently, condemned Russian IHL violations, and noted them in its press release.

Passive Precautions

As I noted in my earlier piece, much of the press release dealt with the issue of passive precautions, a defender’s obligation found in Article 58 of Additional Protocol I (see also Customary International Humanitarian Law study, ch. 6; DoD Law of War Manual, § 5-14).  It requires parties to a conflict to “remove the civilian population, individual civilians and civilian objects under their control from the vicinity of military objectives” and “avoid locating military objectives within or near densely populated areas. Additionally, they must take precautions to protect “the civilian population . . . under their control against the dangers resulting from military operations.” These steps must be taken “to the maximum extent feasible,” a standard that acknowledges the contextual nature of the duty. Curiously, AI did not cite Article 58 in its press release, although it clearly was referring to that obligation. Nor did it cite Article 51(7), which prohibits using the “presence or movements” of civilians to “shield military objectives from attacks or to shield, favor or impede military operations.” This prohibition differs from passive precautions, for it requires intent by the party engaging in the shielding.

The presence of military personnel and equipment near civilians and civilian objects implicates both rules. Concerning shielding, AI did not claim that Ukrainian forces intended to engage in the forbidden tactic, a point the Panel acknowledged. Therefore, it did not examine that issue further. However, as I noted in my earlier post, intent may be inferred from facts, and thus incidents involving co-location should usually be examined for the presence of intent. In this case, the facts contained in the press release do not suggest that Ukrainian forces engaged in intentional shielding, nor can they rule out that possibility.

Compliance with the obligation to take passive precautions is the critical issue. In this regard, the Panel noted that AI documented,

at least 42 specific instances in 19 towns and villages where soldiers were operating in the vicinity of civilians, including 22 incidents involving schools, six involving hospitals, and the rest involving other types of civilian buildings. These incidents included a range of activities, from sleeping and eating to conducting command and control operations to firing weapons systems. In several of these cases, AI determined that attacks by Russian forces that appeared to be targeting the Ukrainian military resulted in death or injury to civilians and damage to civilian objects. (¶ 26).

Importantly, the organization also determined that alternate sites were available to the Ukrainian forces that would have posed less risk to civilians and civilian objects.

There is no prohibition on co-location with civilians and civilians as such. Instead, to breach the obligation of passive precautions by co-location, the party concerned 1) must have available an alternative location that is 2) operationally sensible. Since the Panel appears to have accepted the availability of alternate sites, its assessment hinged on the second prong, military good sense. Key considerations in making such an assessment include the alternative’s impact on mission accomplishment, the risk to own forces, the likelihood that the other location will diminish the risk to civilians and civilian objects, and any effect it might have on other operations (see this HNSJ piece and the DoD Law of War Manual, § 5.2.3.2).

I highlighted a number of factors that might bear on such decisions in my earlier piece, as does the Law of War Manual (e.g., § 5.14.1). Commendably, so too did the Panel.

Such considerations might have included, inter alia, the need to position artillery in a specific location to facilitate a necessary angle of attack on Russian forces; establishing positions that facilitated line of sight connectivity for radar or other communication capabilities or for monitoring enemy positions or forces or gaining access to certain infrastructure such as power or lines of communication necessary to facilitate military operations. Similarly, the operational needs of military units might have required them to billet in a sheltered place with electricity, access to sanitation and potable water, and easy access to roads. (¶ 25).

AI reportedly considered several of these factors but did not include them in the press release. As the Panel points out, it is essential to know if and how the Ukrainian forces assessed their feasibility before drawing any conclusions. It concluded,

Without considering potential justifications for positioning of military forces, and without better understanding and knowledge of whether Ukrainian forces attempted to segregate and protect the civilians in the areas AI examined …, it was not possible for AI to definitively conclude that the Ukrainian military failed to meet its Article 58 AP I obligations.(¶ 30).

The Panel took special note of AI’s concern about the distance of Ukrainian forces from the front lines. In my estimation, AI’s fixation on this point displayed a lack of operational understanding. The Panel came to the same operationally perceptive conclusion: “AI’s determination failed to account for the actual context of the conflict, in which the front lines were often neither narrow nor static due to modern weapon systems with extended reach and high mobility” (¶ 31).

Regarding co-location, the Panel ultimately found that,

the PR’s [press release] insistence that Ukraine had violated its obligations to take passive precautions was too emphatic. Without input and insight from the Ukrainian military concerning whether its troops assessed the feasibility of locating elsewhere and if so, what its feasibility assessments were, legal conclusions should have been caveated appropriately ….  At the very least, given the lack of input from the Ukrainian military, AI should have used more cautious language in the PR, such as noting that Ukraine could or might have violated its obligations under Article 58 AP I and calling for greater scrutiny of the Ukrainian military’s decisions concerning where to locate its forces. (¶ 32).

As will be discussed below, the finding is sound.

Evacuation of Civilians

AI also noted that it was unaware whether the Ukrainian military asked or assisted civilians to evacuate locations near their forces, allegedly “a failure to take all feasible precautions to protect civilians.” I criticized AI’s conclusion on the basis that evacuation is only required if it diminishes the risk to the civilian population and is operationally sensible, a point emphasized in the DoD Law of War Manual (§ 5.14.2). The Panel agreed,

[F]rom a legal perspective, absent engagement on this specific issue with the Ukrainian armed forces, AI could not have definitively concluded that evacuations would have been feasible from a military or, indeed, a humanitarian perspective. An exchange on this issue would have indicated whether evacuating civilians from the buildings in question could have been feasible from a military perspective in the circumstances ruling at the time. Moreover, and very significantly, evacuations might not have been feasible or advisable from a humanitarian perspective. There is no indication of what alternative arrangements for civilians existed or whether forcing them to leave their homes might not have put them in an even more vulnerable situation. There is also no indication in the record of what the specific civilians concerned actually wanted, in particular whether they wanted to leave their homes. (¶ 41).

Warnings

The Panel came to a similar conclusion about warnings. As I explained more fully in my earlier post, Article 58 of Additional Protocol I contains no express obligation to issue warnings. The only such requirement is found in Article 57, which applies to an attacker’s warning to civilians in the target area (so-called “active precautions”). However, I believe a passive precautions duty to warn is a reasonable inference from Article 58(c)’s requirement to take “other necessary precautions to protect the civilian population.” This would include not only warning of enemy attacks but also of its own operations that might put civilians at risk. Of course, this interpretation of the Additional Protocol I obligation, which the Panel labels “progressive” (¶ 49), is subject to a condition of feasibility (including operational considerations like force protection and mission effectiveness). And as I noted previously, the mere presence of the Ukrainian military personnel and equipment may have sufficed to place civilians on notice of the risk, thereby obviating a need for an express warning.

The Panel acknowledged that there might be a duty to warn where “the civilian population is genuinely unaware that they are exposed to an elevated risk of attack by the adversary because defending forces are nearby” (¶ 49). But it correctly emphasized the feasibility condition: “There are many possible military reasons why such specific warnings could not feasibly be given, for example because the defending forces were concerned that civilians might deliberately or unwittingly disclose the location of the military forces and expose them to targeting by posting on (say) social media” (¶ 49). It found that “because IHL does not specifically require defending belligerents to issue such warnings, and in view of the limited information at its disposal, AI’s conclusion that the Ukrainian armed forces violated their obligation to take passive precautions by failing to issue warnings was not sufficiently substantiated” (¶ 51). This finding was, in my estimation, on the mark.

Presence In or Near Schools

In its press release, AI highlighted the presence of Ukrainian troops in and near schools and hospitals. Concerning the former, it noted that the “Ukrainian military has routinely set up bases in schools in towns and villages in Donbas and in the Mykolaiv area. Schools have been temporarily closed to students since the conflict began, but in most cases the buildings were located close to populated civilian neighbourhoods.” While acknowledging that there is no “special protection” in IHL for schools that are not in session, it asserted that “militaries have an obligation to avoid using schools that are near houses or apartment buildings full of civilians, putting these lives at risk, unless there is a compelling military need. If they do so, they should warn civilians and, if necessary, help them evacuate.” In other words, the passive precautions rules apply.

In fact, schools enjoy no “special” protection, whether in session or not. Rather, they are protected by IHL’s rules prohibiting attacks on civilians and civilian objects, forbidding indiscriminate attacks, and requiring compliance with the rule of proportionality and the obligations to take active and passive precautions. The Panel correctly pointed out that by citing a threshold of “compelling military need,” AI arguably set a higher standard than the textual “to the maximum extent feasible” standard (¶ 61). In my view, military need to use a school suffices to allow use. Still, that use immediately triggers the rules just mentioned, the most relevant in this case being passive precautions.

Interestingly, the Panel noted AI’s reference to the Safe Schools Declaration and associated Guidelines for Protecting Schools and Universities from Military Use during Armed Conflict. They are non-binding political commitments that reflect “responsible State behavior.” With respect to the use of abandoned or evacuated schools or universities, Guideline 2 provides that they

should not be used by the fighting forces of parties to armed conflict for any purpose in support of their military effort, except in extenuating circumstances when they are presented with no viable alternative, and only for as long as no choice is possible between such use of the school or university and another feasible method for obtaining a similar military advantage.

AI appears to treat schools that are not in session as analogous to abandoned or evacuated ones, a reasonable conclusion. The alternative would be to treat them as “functioning schools,” which “should not be used by the fighting forces of parties to armed conflict in any way in support of the military effort” (Guideline 1). The Safe Schools declaration has been endorsed by Ukraine but not by Russia (or the United States). Although drawing attention to these non-legally binding norms, AI did not allege their violation.

However, in the Panel’s opinion, the “separate section of the PR on schools nevertheless gives the impression that the use of schools is a special case, which is legally incorrect for the conduct qualified by as IHL violations but might have been justified by the effect of destroying schools on the future right of education” (¶ 62). The Panel did not develop its mention of a human right to education (a legally complex matter legally and factually), nor was it mentioned by AI. Rather, it reemphasized its conclusions as to co-location and feasibility, which, as discussed above, I support fully.

Presence In or Near Medical Facilities

AI also identified five locations where Ukrainian forces were using hospitals as military bases, two areas where soldiers were in the vicinity of hospitals and eating meals there, and one incident of soldiers firing from near a hospital. It stated that the use of hospitals is a “clear violation of international humanitarian law.”

Article 12(4) of Additional Protocol I provides, “Under no circumstances shall medical units be used in an attempt to shield military objectives from attack. Whenever possible, the Parties to the conflict shall ensure that medical units are so sited that attacks against military objectives do not imperil their safety.” Article 18 of Geneva Convention IV less prescriptively states, “In view of the dangers to which hospitals may be exposed by being close to military objectives, it is recommended that such hospitals be situated as far as possible from such objectives.” But as the Panel rightly observed, “Mere presence of members of the armed forces and medical establishments does not violate these rules” (¶  56).

Additionally, the discussion concerning co-location applies equally to this situation. And there is an obligation to “respect and protect” medical facilities (e.g., GC I, art. 19; GC IV, art. 18; AP I, arts. 12-14; Customary International Humanitarian Law Study, Rule 28). However, the protection ceases if medical facilities are used “ to commit, outside their humanitarian function, acts harmful to the enemy,” subject to a warning to desist (GC I, art. 21; GC IV, art. 19; AP I, art. 13; DoD Law of War Manual, § 7.10.3.1). Harmful acts include using the facilities to shelter troops or conducting military operations from them (ICRC Commentaries here, here, and here).

The Panel concluded that it “seems logical that if certain conduct may lead to the loss of protection of hospitals from attacks and a defending party has an obligation to protect hospitals, it may not engage in co-location itself” (¶ 57). Although I did not address this point in my original post, I agree with the Panel’s logic, subject, as always, to a condition of feasibility (as in unavoidable presence during a rapidly evolving troops-in-contact situation). The duty “to respect” further confirms the need to take the passive precautions discussed above, even beyond loss of protection situations. As explained in my original post, it is questionable whether simply being near a medical compound would render the hospital a military objective. However, proximity would certainly trigger passive precautions obligations.

The presence of troops in or near medical facilities is the likeliest basis cited by AI for a Ukrainian IHL violation, especially the alleged use as a location from which to conduct operations. Yet, as the Panel noted, it is difficult to understand the basis for AI’s allegation without an explanation of which IHL rule was violated and how. Thus, I find it appropriate for the Panel to conclude, “AI should either have explained why it considered any presence of soldiers in or near hospitals to be unlawful or should simply have treated those cases in the same manner as other cases of co-location” (¶ 58).

Engagement with Ukraine

Importantly, the Panel determined that AI did not “meaningfully engage” with Ukrainian authorities during the investigation and failed to give it a “right of reply” before publishing its conclusions. As the Panel perceptively noted, “This deprived [AI] of the opportunity to gather information that would have provided a stronger basis for its findings.” Moreover,

in view of the absence of information from the Ukrainian authorities on why members of the armed forces were positioned in civilian areas, the conclusion in the PR that the Ukrainian armed forces had categorically violated the obligation to take passive precautions is insufficiently substantiated. AI would, however, have been justified in concluding that Ukrainian forces could or might have violated IHL even without further potential input from Ukrainian authorities. (¶ 64).

I agree. It cannot be said definitively that Ukraine violated IHL without hearing from Ukrainian authorities. Of course, it is often the case that a party to the conflict refuses to cooperate with such investigations. But failing to exert sufficient efforts to engage with that party, as was apparently the case here, undercuts the credibility of findings. Moreover, absent that attempt, even a suggestion that IHL might have been violated rests on very thin ice.

Concluding Thoughts

From a legal perspective, the most significant point made by the Panel is that,

Without information from the armed forces, it is impossible to categorically determine that a violation of the rules governing the conduct of hostilities has occurred unless there are other clear indications based on patterns of conduct or other circumstantial evidence.

The rules on passive precautions require similar engagement as outlined above, they are not absolute obligations, but instead require belligerents to take precautions to the maximum extent feasible. Feasibility is understood as referring to what is practicable or practically possible, taking into account all of the circumstances ruling at the time, including humanitarian and military considerations. (¶¶ 66-67).

I could not agree more. What the Panel understood, and a point that needed to be more evident in the AI press release, is that IHL is highly contextual. And IHL rules subject to a condition of feasibility sit at the pinnacle of wartime contextuality. Neither the AI press release nor the Panel’s review suffices to rule out the possibility that Ukrainian co-location violated IHL. But the AI press release did not provide the evidentiary basis necessary to conclude that it did.

The Panel must be applauded for producing an objective analysis deeply embedded in the law. As for Amnesty International, I see no reason to deviate from my original assessment that,

Amnesty International has acted irresponsibly by making the claim without providing supporting evidence, citing the specific rules that it believes have been violated, or laying out its legal analysis.These failures have deprived Ukraine of a meaningful opportunity to respond and the international community of an ability to properly assess it.

Fortunately, AI’s establishment of the Panel was responsive to such concerns. But failure to release the Panel’s report when completed, whether in an attempt to bury it or not, was likewise irresponsible; it constitutes a continued failure of leadership. As with the original press release, the delay facilitated Russia’s use of lawfare against Ukraine and unfairly tarnished perceptions of Ukraine’s conduct more broadly. Amnesty International has long contributed measurably to the implementation of IHL and human rights during armed conflict. It can do better than this incident suggests.

***

Michael N. Schmitt is the G. Norman Lieber Distinguished Scholar at the United States Military Academy at West Point. He is also Professor of Public International Law at the University of Reading and Professor Emeritus and Charles H. Stockton Distinguished Scholar-in-Residence at the United States Naval War College.

 

Photo credit: Unsplash