Twitter as Enforcer of the Geneva Conventions

Written by Marko Milanovic

Yesterday Twitter announced a change in its content moderation policies regarding the war in Ukraine. In response to the widespread dissemination of videos and images on the platform of Russian prisoners of war in Ukraine – a reasonably straightforward violation of Article 13(2) of Geneva Convention III, which prohibits exposing PoWs to public curiosity – Twitter adopted the following policy:

In the development and enforcement of our rules, we remain focused on enabling public conversation, and protecting the safety of people both online and offline. We are guided by international humanitarian law, specifically Article 13 of Geneva Convention III (on protecting prisoners of war (PoWs) from any physical or psychological abuse or threat thereof, and encompasses a prohibition on humiliating them) and do not want Twitter to be used by state actors to infringe this law. Our work to protect the conversation is informed by consultation with a wide range of trusted partners, including international human rights organizations, to ensure our approach considers the number of factors at play.

To that end, we will now ask government or state affiliated media accounts to remove any media published that features prisoners of war (PoW) under our private information and media policy.

We will also add a warning interstitial to media published by government or state affiliated media accounts featuring PoWs, that has a compelling public interest. 

Lastly, we will now require the removal of any Tweets, regardless of who posts them, if there is PoW content shared with abusive intent, such as insults, calls for retaliation, mocking/taking pleasure in suffering of PoWs, or for any other behavior that violates the Twitter rules. 

This is clearly the right call, although it will be particularly interesting to observe how Twitter manages any exceptional deviations from the policy on matter of public interest. But it’s even more interesting to observe how a private corporation is here acting as an enforcer of international law against states, and particularly against Ukraine. For years scholars and activists have been arguing that digital platforms should rely on international human rights standards in developing and applying their content moderation policies, in line with the UN Guiding Principles on Business and Human Rights, and this is what the platforms have increasingly been doing. The platforms have also been urged (e.g. notably by David Kaye) to rely on human rights to resist unjustified demands and regulatory interventions by states. They have also acted at least to some extent to prevent and mitigate harms to human rights by states, e.g. through the spread of disinformation.

But here, interestingly, Twitter is acting to enforce IHL, not human rights as such. In doing so it is replicating the structure of the relevant international legal rules, in particular the divide between state and non-state actors. Thus, states are categorically prohibited by Twitter from disseminating content in violation of Article 13(2) GC III, and Twitter’s express basis for doing so is that it does ‘not want Twitter to be used by state actors to infringe this law.’ (Weirdly, this rule becomes a part of Twitter’s private information policy, as if the problem here was an infringement on the human right to privacy and not that of a specific wartime rule of IHL.)

However, when it comes to non-state actors, including private individuals, Twitter will remove such content only if it is shared with abusive intent, in violation of some other rules and policies. The distinction drawn here between state and non-state accounts makes sense in that Article 13 imposes obligations on the states parties to the Conventions and not on other entities, even though media companies clearly may be instrumental in exposing PoWs to public curiosity (note how the ICRC’s new commentary on the Conventions, at paras 1630-1632, is precisely for that reason somewhat coy in (not) saying that the addressees of the relevant obligations are only states). But again it is somewhat strange to say, as Twitter has effectively done, that the videos and images of PoWs contain private information that should not be disclosed on the platform, but only if this is being done by state agents – it’s akin to saying that medical records are private but that Twitter will only remove them from platform if they are being disseminated by state actors, and not by private individuals, or that doxxing is impermissible only if done by states and not by non-state actors. That said, again Twitter clearly did the right thing here (if belatedly), and should be commended for doing so.

Marko Milanovic

Dr Marko Milanovic is Professor of Public International Law at the University of Nottingham School of Law. He is co-editor of EJIL: Talk! and a member of the EJIL's Editorial Board.