Prisoner of War Videos on Facebook

When violent clashes between Armenian and Azerbaijani forces reignited the international armed conflict (IAC) in Nagorno-Karabakh last year, social media was flooded with videos allegedly portraying the mistreatment of prisoners of war (POWs). Recently, Meta, Facebook’s parent company, referred a case concerning one such video to the so-called Oversight Board for review (Article 2 (1) OB Charter). The Board acts as an ‘independent grievance mechanism’ where experts can issue binding decisions or nonbinding policy recommendations on any of Facebook’s content decisions (Article 4 OB Charter), i.e., whether posts or materials can stay on Facebook (a platform with roughly 3 billion active users in 157 countries). While the Oversight Board did not share the video in question, it provided a rough description:

“The video shows people who appear to be Azerbaijani soldiers searching through rubble. The video has been edited so that their faces cannot be seen. They find people in the rubble who are described in the caption as Armenian soldiers. Some appear to be injured, others appear dead. They pull one solider from the rubble, who cries out in pain. His face is visible and he appears injured.”

While the mistreatment of POWs is troubling in itself, this post focuses on the applicability of international humanitarian law (IHL) to Meta’s decision to keep the video on public display even after it became aware of its existence. As a disclaimer, the authors of this post have filed an ‘amicus brief’ on behalf of the Institute for International Law of Peace and Armed Conflict (IFHV) arguing that the video should be taken down. We would like to use this post to discuss some of the conceptual obstacles in applying IHL to foreign corporations’ online conduct during armed conflicts which we could not address in the brief.

First Obstacle: An Incomplete Framework for Applying IHL Online

While the IHL of IACs, i.e., the four Geneva Conventions (GCs) and the first Additional Protocol, is clearly applicable to the situation in Nagorno-Karabakh and the acts portrayed in the video, its applicability to online activities of corporate actors is more contentious.

There is general consensus that IHL applies to “cyber operations executed in the context of an armed conflict” (Tallinn Manual 2.0, p. 375). The term ‘cyber operation’ may include any online activity that falls under IHL’s material scope. Any such cyber operation, however, requires a nexus to the respective armed conflict – the exact nature of which is contested (ibid., p. 376). One camp argues that IHL only covers cyber operations by one conflict party against the other, which would exclude corporate actors who can, by definition, never become party to an IAC (ibid.).

The second position suggests that cyber operations are governed by IHL if “undertaken in furtherance of the hostilities, that is, in order to contribute to the originator’s military effort” (ibid., p. 376). This standard can be read in two ways. The more restrictive reading would require some form of intent to contribute to one side of the hostilities. This interpretation arguably includes private actors conducting cyber operations on the instruction of a conflict party but would not capture Facebook’s more passive role of merely providing an online platform for other actors.

The second interpretation would draw on a matter-of-fact determination of whether any activity, regardless of its originator’s role or intent, actually contributed to the furtherance of the hostilities. Acknowledging states’ primary responsibility to uphold IHL, we argue that the second reading of the nexus requirement ought to be applied in this case because any other understanding would leave online corporate activities that incidentally contribute to the hostilities (e.g., banking, logistics) immune from legal assessment under IHL.

Second Obstacle: No Accountability Framework for Foreign Businesses Operating in Armed Conflicts

While ‘business and human rights’ or corporate criminal liability have received significant attention, the discussion about corporations’ IHL obligations is almost nonexistent apart from private security companies (Doswald-Beck, p. 115 and Chesterman, p. 321). The few notable exceptions (Hughes, p. 47; Bismuth, p. 203 and Aparac, p. 40) have nevertheless advanced two theories for IHL accountability for foreign corporations operating in the context of armed conflicts.

The traditional theory argues that states are under an obligation, by virtue of Common Article 1 of the Geneva Conventions, to “ensure respect” for the rules of IHL in their domestic legal systems (Gillard, p. 130). Since the U.S., where Meta is headquartered, is a party to the Geneva Conventions, all rules contained therein would apply to the conduct at hand, however, only indirectly through U.S. law, which, under the current reading of the Alien Tort Statute, is not enforceable in this case for lack of a minimal connection to the U.S. But one can imagine jurisdictions more willing to hold Meta criminally or civilly accountable for violations of IHL (Mongelard, p. 665).

The second position, tentatively advanced by the International Committee of the Red Cross (ICRC), argues that whenever a corporation’s conduct is “closely linked” to an armed conflict, IHL applies to all corporate activities that might implicate one of its rules. But even this standard seems problematic in relation to Facebook’s passive contribution by hosting content posted by others. In this case, however, Facebook made the conscious decision to keep the video online after review which makes it possible to assert that Facebook itself engaged in an activity that is “closely linked” to the armed conflict.

This, however, leaves us in an uneasy position where everything points in the direction that Meta should follow IHL, but no clear theory that makes IHL directly applicable to incidental foreign corporate conduct has yet emerged.

Third Obstacle: State-Centric Geneva Conventions

Following Meta’s description of the video, the supposed Armenian combatants would qualify as POWs under Article 4 (A) GC III with Azerbaijan as their detaining power (Article 12 GC III).

Article 13 GC III stipulates that “prisoners of war must at all times be protected, particularly […] against insults and public curiosity.” The interpretation of Article 13 has evolved to prohibit any exposure to the public that has the potential to humiliate or identify them. Thus, “the disclosure of photographic and video images […], irrespective of which public communication channel is used, including the internet,” is prohibited (ICRC, p. 592). The reason for this strict prohibition is the aggravated danger of public humiliation of POWs online and the safety risks to their families and themselves (Risius and Meyer, p. 293).

However, the value of video recordings in raising public awareness and as evidence in criminal trials has been cited by Meta as a rationale for keeping the video online. State practice (here, para. 811; here, para. 12.3.3 and here), case law (here), and scholarly opinions (Maia, Kolb and Scalia, pp. 194-200 and Krähenmann, p. 200) support the assertion that a compelling public interest may exceptionally override POWs’ interest in protection against public curiosity. There is agreement however, that a compelling public interest needs to be conclusively demonstrated and that any materials depicting POWs shall not identify or humiliate them, except for cases of utmost public importance (e.g., a senior government official POW is wanted). The commentary to GCIII, therefore, recommends that “the media should always resort to appropriate methods, such as blurring, pixelating or otherwise obscuring faces and name tags, altering voices or filming from a certain distance, in order to serve their function without disclosing the prisoners’ identities” (ICRC, p. 594).

In this case, the face of at least one POW was visible, possibly exposing the injured individual to identification and humiliation. Meta has argued “that the public interest in seeing the content outweighed the risk to the safety and dignity of the prisoners of war.” But the public’s rather abstract “general interest in seeing the content” does not satisfy the specificity required under the “compelling public interest” standard. Meta has also not balanced the public’s interest against the POWs’ right to not be identified or humiliated. Therefore, Meta’s decision to keep this video publicly available is inconsistent with Article 13 GC III.

But, once again, we run into the problem that Article 13 GC III specifically covers “any unlawful act or omission by the Detaining Power” suggesting that the conflict parties have the sole responsibility to protect POWs (Article 12 GC III). If one follows this systematic, Facebook would be off the hook. One could only blame Azerbaijan for its soldiers’ violation of Article 13 GC III.

Where Do We Go From Here?

Reflecting on our brief in this case, we realized that IHL can be a powerful tool for evaluating the online conduct of foreign corporate actors in armed conflicts. Meta, Twitter, and others even embrace the language of human rights and IHL in their policy documents. Our problem is, however, that the entire IHL framework depends on the assumption that all conduct in the context of an IAC is connected to a (state) conflict party, especially in relation to the treatment of POWs. This ignores the new reality that transnational corporations play an important role in online warfare and are often beyond the control of the conflict parties. Perhaps the time has come to update the ICRC’s 2006 guidelines on ‘Business and IHL’ to translate the state-centric IHL language into a concrete code of conduct for businesses and emphasize that armed conflicts are not a legal vacuum for corporations.

The “Bofaxe” series appears as part of a collaboration between the IFHV and Völkerrechtsblog.