Propaganda and Accountability for International Crimes in the Age of Social Media: Revisiting Accomplice Liability in International Criminal Law

[Talita de Souza Dias is a DPhil Candidate and a Tutor in Public International Law and International Criminal Law at the Faculty of Law, University of Oxford.]

In the past few weeks, there have been a series of news reports on the role that data-mining firm Cambridge Analytica played in the outcome of the 2016 US elections, which led to Donald Trump’s controversial victory (see, e.g., here and here). In essence, the corporation is being accused of manipulating the results of that election, especially by harvesting protected personal data from Facebook and by making targeted posts on this and other social media platforms, including with the aim of discrediting other candidates and spreading fake news. Significantly, there have also been allegations that Cambridge Analytica has been employing similar marketing tactics in Kenya for the purposes of manipulating the country’s 2013 presidential elections, and the primary elections that took place earlier this year, in favour of Kenya’s President Uhuru Kenyatta (see here, here and here). According to the company’s website, it collected data on the ‘electorate’s real needs (jobs) and fears (tribal violence)’ to ‘draft an effective campaign strategy’. Footage obtained from a hidden camera also shows one of Cambridge Analytica’s managing directors admitting to the company’s role in rebranding Kenyatta’s entire party twice, such as by drafting speeches and manifestos, managing their media campaign, and ‘stag(ing) the whole thing’. In light of these allegations, and Kenya’s recent history of post-election violence, there have been fears that the company’s marketing strategies might contribute to yet another wave of ethnic hatred and physical confrontation in the country (see here). It is important to recall that Kenya’s 2007 post-election violence gave rise to an investigation and a series of prosecutions before the International Criminal Court (ICC), including against Kenyatta.

To be sure, war propaganda and the involvement of the media in the commission of international crimes are not new phenomena, as the examples of Nazi Germany and Rwanda testify. However, the use of social media adds a new lawyer of complexity and raises a series of new questions on the international criminal responsibility of the individuals involved in marketing campaigns such as the ones orchestrated by Cambridge Analytica. Two such questions are of particular relevance. First, is the contribution to an election campaign, or the manipulation of an election sufficient to engage the individual criminal reasonability of social media operators for any ensuing international crimes? Secondly, in what other ways the use of social media can give rise to international criminal responsibility?

Since we are talking about the use of social media to assist with or contribute to the commission of international crimes by other individuals, it is accomplice or accessorial liability that is most relevant to our scenarios. In particular, there are two key modes of accomplice liability that could potentially fit the factual pattern at hand, namely, aiding and abetting, or instigating, soliciting, inducing and inciting (or simply ‘instigation’). Under customary international law (CIL) and in the Rome Statute of the ICC, the objective and subjective elements of instigation have not been subject to any significant controversy. Indeed, under both CIL and the Rome Statute, instigation requires conduct that prompts, drives or causes the principal perpetrator to commit the crime. Thus, the instigation must be a ‘clear’ or ‘direct’ contributing factor to the commission of the crime. This does not mean that the contribution must be essential to the crime, but it must have substantially contributed to it. As to the mental element, the default standard in either regime seems to apply: under CIL, intentional conduct must be coupled with knowledge of the substantial likelihood of the crime (i.e. recklessness), whereas in the Rome Statute Article 30 requires intent both in relation to the conduct and the result (i.e. intention or dolus directus).

The picture is somewhat blurrier when it comes to aiding and abetting. On the one hand, the objective elements of this mode of liability seem to be same under CIL and the Rome Statute: although a substantial or significant contribution is required, this broadly includes any contribution that is more than a de minimus. Thus, what separates an inadvertent contribution from aiding and abetting is the mental element. Again, the mental element in the Rome Statute differs from the one under CIL. After some confusion in the case-law of the ICTY as to whether or not ‘specific direction’ was required for aiding and abetting, it now appears to be settled that, under CIL, knowledge is the requisite mental element for that mode of liability. Thus, the accused must know that his/her conduct assists, or is substantially likely to assist a specific crime. As to the Rome Statute, Article 25(3)(c) requires the aider to have ‘the purpose of facilitating the commission of [the] crime’. Thus, the ICC has held that Article 30’s default requirement of intent also applies to aiding and abetting (see, for instance, the recent Appeal Judgement in the Bemba et al. case, paras 1390-1391, 1399-1401). Nonetheless, in an attempt to address neutral or non-criminal types of assistance, some commentators have proposed a restrictive interpretation of aiding and abetting (see, e.g. Kai Ambos’ post on the evacuation of civilians in Syria and the Separate Opinion of Judge Fernández de Gurmendi in the Mbarushimana Appeals Confirmation Decision). The idea is to require, aside from a minimum and purposeful contribution, some normative nexus between the aider and the principal’s conduct. Although it is not very clear what such normative standard entails, it appears that it is mainly based on the unlawfulness of the assistance and on the fact that it increases the risk of the commission of the crime. The same interpretation has been applied to the Rome Statute’s residual mode of accomplice liability under Article 25(3)(d). However, this provision seems to be less demanding than Article 25(3)(c), in that it only requires the accomplice to have knowledge that a certain group has the intention or common purpose to commit a crime.

Applying this law to our relevant factual scenarios, it seems that assistance to the election of an individual who turns out to be a perpetrator of international crimes, by means involving the use of social media, could potentially amount to instigation or aiding and abetting under CIL and Article 25(3)(b) and (c) of the Rome Statute, or to the residual mode of liability under Article 25(3)(d) of the Statute. This is because, as the example of Cambridge Analytica shows, social media can be a powerful if not decisive tool for putting such individuals in a position to commit international crimes, such as crimes against humanity, genocide, war crimes and the crime of aggression. It can also shape their political agendas and discourse, and so directly influence the commission of such crimes. However, the accomplice must either act with knowledge of the principal’s intent to commit the crime(s), or share the latter’s intent. In addition, if one adopts a restrictive interpretation of accomplice liability, the lawfulness of the accomplice’s contribution would be a decisive factor. In this regard, there seems to be general agreement that the use of unauthorised personal data is unlawful in most domestic legal systems and under international law (it is at the very least a breach of the human right to privacy). However, it remains open to question whether targeted posts which influence the behaviour of the electorate are unlawful under either body of law. Although it is likely that derogatory posts on social media can amount to an unlawful contribution (as these may constitute defamation, slander or libel in many domestic legal systems), it is unclear what other types of posts can have the same legal qualification.

Regarding other types of assistance, it seems that social media has countless ways of influencing or contributing to the commission of international crimes. In particular, following the footsteps of Nazi war propaganda and RTL’s radio broadcasts in Rwanda, social media can be used to instigate or incite the commission of international crimes by the masses. In addition, data-mining or harvesting can also be used in the planning or preparation of international crimes, especially those targeting predominantly civilian populations, such as crimes against humanity and genocide. If it is proved that such acts substantially contribute to those crimes and that the accomplice acts with the requisite knowledge or intent, then accessorial liability could be engaged under CIL and the Rome Statute, as instigation, aiding and abetting, or the ICC’s residual mode of liability.

Furthermore, aside from accessorial liability for the acts of others, social media operators could be held responsible for their own acts, as principal perpetrators. This could be the case if such acts consist of direct and public incitement to commit genocide, or hate speech as an underlying act of crime against humanity (if made in conjunction with other acts that severely deprive individuals of their human rights). Their acts could also fall under superior responsibility of civilians under CIL or Article 28(b) of the Rome Statute, provided that they have effective control over social media platforms, actual or constructive knowledge of the commission of international crimes by their subordinates, and fail to prevent or punish such crimes.

In conclusion, the use of social media to manipulate elections and to provide other types of assistance to international crimes can potentially give rise to individual criminal responsibility under international law. In addition to the contribution itself, other objective and mental elements must still be proved, but such responsibility cannot be ruled out. Importantly, the ICC still has jurisdiction of over the situation in Kenya, and it may well re-open its investigation and prosecutions on this matter. It also has jurisdiction over similar scenarios taking place on the territory of its states parties and acts committed by nationals of such states, at the very least. In addition, domestic courts also have jurisdiction over similar facts, including under the principle of universality, at least when they amount to crimes against humanity, genocide and war crimes.