I. Introduction
On August 25, 2017, in the western region of Myanmar known as Rakhine State, a conflict of untold horror—long in the making—spiraled out of control.1 In the day’s first hours, a band of Rohingya Islamist insurgents, armed with sticks, knives, and makeshift bombs attacked a series of government outposts.2 The attacks left seventy-one dead, among them fifty-nine Rohingya militants and twelve government officials.3 In response, the Myanmar military, known as the Tatmadaw, announced the launch of “clearance operations,” an all-out assault on thousands of Rohingya civilians.4 The Tatmadaw razed whole villages, killing thousands of people.5 Women and girls were subjected to brutal gang rapes.6 Homes were reduced to ash.7 Altogether, the operations forced more than 745,000 Rohingya to flee to Bangladesh.8 According to the U.N., over 600,000 Rohingya remain at “serious risk of genocide” in Myanmar.9
The sheer scale of violence led journalists, human rights organizations, and the U.N. to investigate the Tatmadaw’s response to the August 25 incident, including how an entire nation remained passive as thousands of Rohingya civilians were killed, raped, and displaced from their homes.10 These investigations revealed an elaborate disinformation campaign, spanning years, whereby senior Tatmadaw officials used Facebook, the “main mode of communication among the public,” to fuel nationwide majoritarian hatred toward the Rohingya.11 All the while, Facebook provided the Tatmadaw with a massive platform to propagate dehumanizing narratives about the Rohingya.12 For example, officials used their accounts to frame the Rohingya as illegal immigrants, terrorists, and “inherently violent.”13 As the Tatmadaw used Facebook to disparage the minority group, civilian users joined in the attacks online.14 Reuters uncovered more than 1,000 examples of violent Facebook posts, some dating back to 2012, which referred to the Rohingya as “dogs,” “maggots,” and “rapists” who ought to be “fed to pigs” and “exterminated.”15
As dehumanizing speech spread on Facebook, and the hostility toward the Rohingya worsened, human rights activists and experts on the region reportedly warned Facebook executives as early as 2013 that their platform facilitated and exacerbated the ethnic tensions in Myanmar.16 Despite these warnings, Facebook continued providing Tatmadaw officials with a platform, devoting insufficient resources toward preventing the spread of disinformation. For example, around 2015, “there were only two people at Facebook who could speak Burmese.”17 It was not until 2018, when U.N. investigators formally accused Tatmadaw officials of genocide, that the company began the systematic removal of content and accounts belonging to military leaders.18 Facebook thereafter conceded in a blog post, “[w]e weren’t doing enough to help prevent our platform from being used to foment division and incite offline violence.”19
The U.N. concluded, after lengthy investigation, that there was “sufficient information to warrant the investigation and prosecution” of senior Tatmadaw officials for genocide.20 For its part, the International Criminal Court (ICC) held that although Myanmar is not a member of the court, the ICC may exercise jurisdiction over crimes committed against the Rohingya with a “cross-border” nature.21 As of June 2019, the ICC prosecutor was considering the launch of a formal investigation into the actions of Tatmadaw officials.22 But even as international courts and organizations contemplate whether the Tatmadaw committed genocide, there is confusion around how to address Facebook’s involvement in the massacre. The U.N. identifies the company’s role as “significant,” noting in a fact-finding report that the platform “has been a useful instrument for those seeking to spread hate” and that Facebook’s response “has been slow and ineffective.”23 Nonetheless, in the same report, investigators did not identify Facebook as criminally responsible.24
The confusion around Facebook’s role in Myanmar raises broader questions about whether social media companies risk international criminal liability when their platforms are used by bad actors to incite offline violence—and more specifically, genocide. These questions are of increasing importance, as up to seventy governments and political parties across the globe actively use social media to spread disinformation.25 Notwithstanding the myriad of procedural obstacles (for example, exercising jurisdiction) that stand in the way of prosecution, this Comment focuses on the more narrow, substantive question of whether social media companies can be criminally complicit in direct and public incitement to genocide.26 This Comment concludes social media companies can be criminally complicit in incitement.
In arguing that social media companies may risk international criminal liability where their platforms are used to incite genocide, this Comment proceeds in four parts. Section II discusses the substantive international crimes of genocide and direct and public incitement to commit genocide. This analysis focuses primarily on the latter crime of incitement but nonetheless discusses genocide to shed light on the basic principles underlying incitement. In Section III, this Comment identifies relevant modes of liability, including aiding and abetting and common purpose liability. These modes of liability stipulate the requirements which, if satisfied, would attach criminal responsibility to social media officials for the commission of a substantive offense. Section IV analyzes how contemporary legal standards regarding corporate liability limit culpability to natural persons. Having introduced the pertinent international criminal law (ICL), Section V applies the law to hypothetical cases at different levels of the corporate hierarchy.
II. Pertinent International Crimes
TOPThis Comment opens with a discussion of two international crimes: genocide and direct and public incitement to commit genocide. For the scope of this Comment, ICL refers to the body of international law, composed largely of treaty law and caselaw, which imposes criminal liability on individuals.27 As the most recent and comprehensive treaty on ICL, the Rome Statute of the International Criminal Court (Rome Statute) necessarily anchors discussion of ICL. Entered into force in 2002, the Rome Statute established the ICC as a permanent international court and empowered the court to examine ICL principles beyond the treaty’s plain terms.28 Article 21 of the Rome Statute authorizes the ICC to apply ICL from “applicable treaties and the principles and rules of international law.”29 Accordingly, a proper understanding of ICL today requires reference to older treaty law as well as caselaw more generally.
This Comment ultimately focuses on direct and public incitement to commit genocide in the context of social media. However, a brief overview of genocide provides useful context to understanding incitement.
Considered by some to be the “crime of crimes,” genocide carries unique weight in ICL.30 Genocide was criminalized in 1951, when the Convention on the Prevention and Punishment of Genocide (Genocide Convention) entered into force.31 The treaty’s ratification came as the world reeled from the horrors that World War II inflicted upon civilian populations. Mass atrocities committed against the Jewish people at the hands of the Nazi Party shocked a universal conscience.32
As the International Military Tribunal (IMT) sought to prosecute the Nazis for their crimes against humanity, scholars and world leaders coalesced around a particular theory of “genocide.” Raphael Lemkin, a Polish-Jewish lawyer and the Genocide Convention’s primary drafter, coined the phrase genocide “combining geno-, from the Greek word for race or tribe, with -cide, from the Latin word for killing.”33 Matthew Lippman writes that “[a]ccording to Lemkin, genocide involved a two phase process, the destruction of the ‘national pattern of the oppressed group’ and the ‘imposition of the national pattern of the oppressor.’”34 Genocide was thus conceived as a system-wide offense, a sociological restructuring whereby institutions of power, ranging from government to media, are weaponized to will the destruction of a particular group over time and in multiple stages.
In converting genocide from a theory to a crime, the international community understood the magnitude of the offense. The Genocide Convention preamble captured a universal sentiment, characterizing genocide as a crime that “inflicted great losses on humanity” and demanding international cooperation “to liberate mankind from such an odious scourge.”35 The parties criminalized genocide in Articles I and II of the Genocide Convention, which provide:
Article I: The Contracting Parties confirm that genocide, whether committed in time of peace or in time of war, is a crime under international law which they undertake to prevent and to punish.
Article II: In the present Convention, genocide means any of the following acts committed with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such:
(a) Killing members of the group;
(b) Causing serious bodily or mental harm to members of the group;
(c) Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part;
(d) Imposing measures intended to prevent births within the group;
(e) Forcibly transferring children of the group to another group.36
As ICL developed, genocide’s status as a crime became unequivocal. In the Statutes of the International Criminal Tribunal for the former Yugoslavia (ICTY) and the International Criminal Tribunal for Rwanda (ICTR)—Articles 4(2) and 2(2), respectively—the international community reaffirmed genocide as a crime, using the exact language of the Genocide Convention.37 The Rome Statute continued this tradition in Article 6 of that instrument, further cementing global fidelity toward genocide’s prevention.38
While treaties criminalized genocide, international courts expanded on the crime’s elements through interpretation, as discussed by Grant Dawson and Rachel Boynton.39 For example, international courts have interpreted “killing members of the group” as murder, excluding non-intentional homicides.40 The meaning of “causing serious bodily or mental harm” has been determined on a case-by-case basis, with consideration made to the particular circumstances.41 International courts have found this crime to include torture, sexual violence and rape, degrading treatment, threats of death, and “harm that damages health or causes disfigurement or injury.”42 “Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part” has been extended beyond mere killing or physical harm to include instances where a protected group is subjected to the “systematic expulsion from homes.”43 The remaining genocidal acts, including the imposition of measures intended to prevent birth and the forcible transfer of children, have been less developed by courts.44
In addition to the foregoing genocidal acts (actus reus), courts have held the mens rea of genocide to be purposeful.45 Courts have also required satisfaction of specific intent, that is, the intent to destroy, in whole or in part, a protected group (a national, ethnical, racial, or religious group).46 Where there is an absence of direct evidence, specific intent may be inferred from the circumstances.47 A defendant is therefore guilty of genocide if he intentionally (mens rea) commits a genocidal act (actus reus) with the specific intent to destroy, in whole or in part, a protected group.
B. Incitement to Genocide
TOPIn addition to genocide, international law has criminalized direct and public incitement to commit genocide. However, because the Rome Statute lists incitement in Article 25 entitled “Individual criminal responsibility” and not in Article 5 entitled “Crimes within the jurisdiction of the Court,” doubts have been raised as to whether incitement is a crime (a substantive offense independent from genocide) or a mode of liability (a means by which liability for genocide attaches).48 Notwithstanding Article 25 placement, this Comment concludes that incitement can and should be treated as a crime, not a mode of liability. History, treaty law, and relevant precedent all point towards such a reading.
1. Treaty law: the Genocide Convention, the ICTY and ICTR Statutes, and the Rome Statute
The criminalization of incitement to genocide reflects the longstanding view that genocide is a multi-stage process which begins long before systemic violence occurs. Lemkin saw this process as one moving “from stigmatisation and dehumanisation through violence and terror and eventual annihilation.”49 Accordingly, to curb genocide as early as possible, the Genocide Convention criminalized not only genocidal acts, which manifest in the latter stages of the crime, but also other acts that may occur earlier.50 Those additional crimes, stipulated in Article III, are directed towards the prevention of “stigmatisation” and “dehumanisation.” Included among these Article III crimes is the “[d]irect and public incitement to commit genocide.”51
Following the Genocide Convention, the ICTY and ICTR Statutes similarly established incitement as a crime, using the same language as Article III of the Genocide Convention. The ICTY Statute criminalizes incitement in Article 4(3)(c) and the ICTR Statute does so in Article 2(3)(c).52 Complicating the picture with respect to treaty law, however, is the Rome Statute’s treatment of incitement.53
Unlike the Genocide Convention and the ICTY and ICTR Statutes, the Rome Statute does not adopt the exact language of the previous treaties on incitement. Article 25 of the Rome Statute states, “In accordance with this Statute, a person shall be criminally responsible and liable for punishment for a crime within the jurisdiction of the Court if that person . . . [i]n respect of the crime of genocide, directly and publicly incites others to commit genocide.”54 Article 25 is entitled “Individual criminal responsibility” and outlines the rules around various modes of liability.55
Notwithstanding incitement’s placement in Article 25, the Commentary on the Law of the International Criminal Court (CLICC) takes the position that incitement to genocide is a crime under the Rome Statute, not a mode of liability. CLICC editor Mark Klamberg writes;
Article 25(3)(e) of the ICC Statute criminalises direct and public incitement of others to commit genocide. It is in substance identical to Article III(c) of the 1948 Convention on the Prevention and Punishment of the Crime of Genocide, and the ICTY and ICTR Statutes. Genocide is the only international crime to which public incitement has been criminalised. The reason for this provision is to prevent the early stages of genocide even prior to the preparation or attempt thereof.56
W.K. Timmermann concurs with this position, writing that incitement “has of course been unequivocally recognized” as a crime.57
At least one commentator has contended that incitement’s inclusion in Article 25 converted it from a crime to a mode of liability.58 Thomas Davies argues the Rome Statute “makes a crucial departure from earlier instruments with respect to genocide” and “denies incitement the status of an independent crime, and instead presents it as a type of individual criminal responsibility for genocide.”59 As Davies notes, Article 5 is entitled “Crimes within the jurisdiction of the Court” and makes no mention of incitement.60 This approach finds further support in Article 22(2), which provides that “[t]he definition of a crime shall be strictly construed” and that ambiguities should be “interpreted in favour of the person being investigated, prosecuted or convicted.”61
The implications of this approach are significant. If Davies is correct, the Rome Statute substantially reduces the scope of criminal conduct under ICL.62 For example, if direct and public incitement is a crime, it creates “various forms of secondary liability to hold a range of individuals responsible beyond those who directly commit the incitement.”63 Davies points to examples of who might be prosecuted for complicity in incitement—the speechwriter who pens a genocide-inciting speech or the manager of a radio station that airs inciting broadcasts.64 But if incitement is not a crime, it cannot generate secondary liability. Those who assist inciters would be immune from prosecution. Davies recognizes this as a drawback. To “correct the problem,” he recommends the Rome Statute be amended.65
While an amendment would satisfy a strict textualist approach to the Rome Statute, the treaty as currently written can and should be interpreted as criminalizing incitement to genocide. Article 21 entitled “Applicable law” requires the ICC to first apply the Rome Statute, but it also forecloses disregard for broader principles of international law.66 Specifically, upon review of the Rome Statute, Article 21(1)(b) requires the ICC to apply, where appropriate, “applicable treaties and the principles and rules of international law.”67 One such rule, already applied by ICC judges, is the “General Rule of Interpretation” codified in Article 31 of the Vienna Convention on the Law of Treaties.68 Article 31 states, “A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in light of its object and purpose.”69 Article 31 further notes that a treaty’s preamble should be considered as part of an object and purpose inquiry.70
The preamble to the Rome Statute establishes that the object and purpose of the treaty was to ensure the “effective prosecution” of “the most serious crimes of concern to the international community as a whole” and “to contribute to the prevention of such crimes.”71 The States Parties reflected on the 20th century, considering themselves “[m]indful that during this century millions of children, women and men have been victims of unimaginable atrocities that deeply shock the conscience of humanity.”72 As evidenced by the relevant treaty law and caselaw, incitement to genocide—by virtue of its close nexus to genocide—has been consistently considered among the world’s most serious crimes.73 The entire point of making incitement itself a crime was to prevent genocide, and such prevention is explicitly contemplated in the Rome Statute’s preamble.74 It therefore makes little sense to halt an analysis of incitement at its location in the Rome Statute and the text of article titles. Rather, the inclusion of incitement at all demonstrates that the international community sought to continue its punishment, consistent with decades of well-developed international law.75
Article 22(2)’s requirement that ambiguities be resolved in favor of the defendant does not lead to a different result.76 Any ambiguity resulting from incitement’s placement in Article 25 is clarified by application of the General Rule of Interpretation, which is not only permitted but also compelled by Article 21. Accordingly, this Comment adopts the CLICC approach to the Rome Statute—incitement is a crime not a mode of liability.
2. Caselaw: the International Military Tribunal and the ICTR
In addition to treaty law, caselaw confirms that incitement to genocide is a crime.77 Where treaties fail to expand on incitement’s elements, precedent provides an indispensable tool to understand incitement doctrine. Specifically, decisions by the IMT and the ICTR reveal that in order for the defendant to be convicted of incitement to genocide, the following elements must be satisfied: (1) the incitement must be intentional; (2) it must be public; (3) it must be direct; and (4) the inciter must have the specific intent to cause genocide.78 The seminal cases illustrate how incitement’s criminalization serves to prevent the “stigmatisation” and “dehumanisation” that compose the early stages of genocide.79
a) IMT Caselaw
Two IMT cases, Streicher and Fritzche, provide an early window into incitement doctrine.80 They were decided after WWII, prior to the ratification of the Genocide Convention. Accordingly, the prosecution did not seek convictions for “incitement to genocide” but rather for crimes against humanity.81 Nonetheless, these cases established the foundations of incitement upon which the ICTR later built.
In Streicher, defendant Julius Streicher was the editor of an anti-Semitic German weekly newspaper, Der Stürmer, a publication which in 1935 had a circulation of 600,000.82 From 1923 to 1945, he played a role in the routine dissemination of content which encouraged violence towards the Jewish people.83 For example, in twenty-three different articles, Streicher called for their extermination “root and branch.”84 He commonly used dehumanizing phrases such as “germ,” “pest,” or “parasite” when referring to Jews.85 In 1940, Streicher published a letter from one of his readers that “compared Jews with swarms of locusts which must be exterminated completely.”86 Streicher continued publishing propaganda even as the mass execution of Jews was ongoing.87 The IMT concluded that “[i]n his speeches and articles, week after week, month after month, [Streicher] infected the German mind with the virus of anti-Semitism, and incited the German people to active persecution.”88 Streicher was convicted for crimes against humanity and subsequently put to death.89
In Fritzsche, the defendant Hans Fritzsche held various roles in German media, including as a radio commentator, as the chief of the Home Press Division (a propaganda news service), and eventually as the head of the Radio Division of the German Propaganda Ministry.90 Fritzsche oversaw the publication of 2,300 German daily newspapers, a role which involved issuing media guidance—developed at higher levels of the Nazi bureaucracy—highlighting themes such as “the Jewish problem” and “the problem of living space.”91 As the leader of the Radio Division, Fritzsche came under the supervision of Joseph Goebbels and would relay the “news” of the day, which was often false.92 In contrast to Streicher, the IMT did not find Fritzsche guilty of incitement, reasoning that although he “sometimes made strong statements of a propagandistic nature in his broadcasts” and although his speeches showed “definite anti-Semitism,” the prosecution failed to show that his statements “were intended to incite the German people to commit atrocities.”93 The IMT also made note of the fact that Fritzsche’s “position and official duties were not sufficiently important . . . to infer that he took part in originating or formulating propaganda campaigns.”94
The IMT cases demonstrate how courts reviewed not only the content of potentially inciting statements, to determine whether they amounted to calls for genocide, but also whether they were broadcast to substantial audiences (a foreshadowing of the eventual “public” requirement). Perhaps the most important contribution of these cases, however, is the special attention paid by the IMT to whether the statements were part of a deliberate campaign, suggesting that even where there is no explicit call for genocide, a systematic propaganda campaign may rise to the level of incitement (a foreshadowing of the eventual “direct” requirement). This construction of incitement is consistent with Lemkin’s theory of genocide as an attempt to fundamentally restructure society. Such principles informed the ICTR’s analysis and expansion of incitement doctrine.
b) ICTR Caselaw
The ICTR built on the IMT’s jurisprudence in its own series of cases. Akayesu was the first such case.95 The defendant Jean-Paul Akayesu was the mayor of the Rwandan town of Taba.96 He addressed a crowd of over 100 people, calling on them to “unite in order to eliminate what he termed the sole enemy: the accomplices of the Inkotanyi.”97 Akayesu also read aloud the names of those he purported to be accomplices.98 While the Inkotanyi was a particular military-political faction of the Tutsis, the ICTR found, after hearing testimony on cultural and linguistic context, that “Akayesu himself was fully aware . . . that his call to fight against the accomplices of the Inkotanyi would be construed as a call to kill the Tutsi in general.”99 Akayesu was convicted of the crime of direct and public incitement to commit genocide, which the ICTR noted to be “distinct from the crime of genocide.”100
Importantly, in the Akayesu decision, the ICTR also elaborated the definition of incitement:
[D]irect and public incitement must be defined for the purposes of interpreting Article 2(3)(c), as directly provoking the perpetrator(s) to commit genocide, whether through speeches, shouting or threats uttered in public places or at public gatherings, or through the sale or dissemination, offer for sale or display of written material or printed matter in public places or at public gatherings, or through the public display of placards or posters, or through any other means of audiovisual communication.101
The Court further held that directness must be considered “in light of its cultural and linguistic content,” noting that an implicit statement, like the one made by Akayesu, may be sufficient.102
While Akayesu established that the ICTR would punish incitement as a crime, the communication in that case was less analogous to the widely circulated publications in Streicher, where the incitements were part of a systematic campaign. On the other hand, in Nahimana, known widely as the Media Case, the ICTR considered the issue of national propaganda campaigns.
The Media Case involved three defendants, including Ferdinand Nahimana, Jean Bosco Barayagwiza, and Hassan Ngeze—the purported “masterminds behind a media campaign to desensitize the Hutu population and incite them to murder the Tutsi population in Rwanda in 1994.”103 Nahimana and Barayagwiza together founded Radio Télévision Libre des Mille Collines (RTLM) and from July 1993 to July 1994 broadcast anti-Tutsi messages to nationwide audiences.104 The ICTR Trial Chamber convicted them of direct and public incitement to commit genocide, noting that “RTLM broadcasts engaged in ethnic stereotyping in a manner that promoted contempt and hatred for the Tutsi population and called on listeners to seek out and take up arms against the enemy”105 and that “Nahimana and Barayagwiza were, respectively, ‘number one’ and ‘number two’ in the top management of the radio.”106 Ngeze similarly managed a newsletter called Kangura, which from 1990 to 1995 published articles that also conveyed “hate-filled messages” about the Tutsis.107 The Trial Chamber accordingly convicted him of incitement.108
Susan Benesch notes how the ICTR Trial Chamber’s decision failed to specify which acts constituted “incitement to genocide.”109 Moreover, the Trial Chamber’s decision aroused concerns that incitement had been erroneously conflated with hate speech.110 The ICTR Appeals Chamber sought to resolve these problems through a meticulous analysis of each RTLM broadcast and Kangura article, ultimately concluding that the Trial Chamber had not confused hate speech with incitement to genocide.111 Instead, the Appeals Chamber affirmed the Trial Chamber’s holding that an incitement need not be explicit, reasoning that cultural context may demonstrate that an audience clearly understood the statements as inciting genocide.112 The Appeals Chamber did, however, reverse Barayagwiza’s incitement conviction, noting—like the IMT in Fritzsche—that he was Nahimana’s subordinate. Nahimana and Ngeze’s convictions were affirmed.113
The foregoing IMT and ICTR cases provide rich guidance as to incitement’s four elements, particularly the harder-to-prove elements of “direct” and “public.” Public incitement means “that the call for criminal action is communicated to a number of individuals in a public place or to members of the general public at large particularly by technological means of mass communication, such as by radio or by television.”114 Direct incitement may be either a particular statement that explicitly calls for genocide or implicit statements that, in their cultural context, are understood as calls for genocide.115 International courts especially consider systematic propaganda campaigns, in addition to individual statements, as satisfying incitement; however, in analyzing campaigns they are less likely to convict the propagandist’s deputies.
III. Modes of Liability
TOPIn addition to the principal perpetrator of a crime, ICL allows for the prosecution of other actors for the same crime through various modes of liability. This Section examines two modes which may impose liability on the employees of social media companies, aiding and abetting liability and common purpose liability. Aiding and abetting extends criminal liability to those who assist in a crime. Common purpose liability—a relatively new form of liability created by the Rome Statute—functions similarly to aiding and abetting, when the crime has been committed by a group. This Comment therefore considers both modes to be variations of complicity. Article 25 of the Rome Statute explicates their requirements, which control for the purpose of ICC jurisprudence. After discussing aiding and abetting and common purpose liability, this Section examines complicity in incitement.
A. Aiding and Abetting
TOPAiding and abetting liability has long been a feature of ICL.116 According to Doug Cassel, the doctrine dates back to the 1945 Charter of the International Military Tribunal, which imposed liability on “accomplices participating in the formulation or execution of a common plan or conspiracy to commit” a crime listed in the Charter.117 Aiding and abetting is also captured in Article 25(3)(c) of the Rome Statute which states,
In accordance with this Statute, a person shall be criminally responsible and liable for punishment for a crime within the jurisdiction of the Court if that person . . . [f]or the purpose of facilitating the commission of such a crime, aids, abets or otherwise assists in the commission or its attempted commission, including providing the means for its commission.118
Pursuant to Article 25(3)(c), aiding and abetting consists of two elements, the actus reus and the mens rea.119
The ICC construes the actus reus of aiding and abetting to be the provision of practical or material as well as moral or psychological assistance to the principal perpetrator.120 However, the “precise actus reus threshold” remains an open question.121 The tribunals may provide useful insight to this effect. They have held that “aiding and abetting requires acts or omissions that assist, encourage or lend moral support to crimes.”122 The tribunals have also required that the aider and abettor’s conduct “substantially” contribute to the commission, similar to the ICC’s notion of “material” assistance.123
The mens rea for aiding and abetting under the Rome Statute, as indicated in Article 25(3)(c), is purpose.124 This requirement is markedly harder to prove than the mens rea applied by ad hoc tribunals, which is knowledge.125 For example, the ICTY Trial Chamber in Šešelj held that, as it pertains to the defendant’s contributions, “the aider and abettor must have known that these acts had contributed to the perpetration of the crime and been aware of the essential elements of the crime, including the intent of the principal perpetrator, without necessarily knowing the exact crime that was intended or committed.”126 Although the Rome Statute’s mens rea for aiding and abetting is higher than the standard applied by the ad hoc tribunals, common purpose liability—a Rome Statute innovation—effectively lowers it back to knowledge for group crimes.
Article 25(3)(d) of the Rome Statute establishes “common purpose liability,” which this Comment considers a form of complicity. The Article stipulates:
In accordance with this Statute, a person shall be criminally responsible and liable for punishment for a crime within the jurisdiction of the Court if that person . . . in any other way contributes to the commission or attempted commission of such a crime by a group of persons acting with a common purpose. Such contribution shall be intentional and shall either:
(i) Be made with the aim of furthering the criminal activity or criminal purpose of the group, where such activity or purpose involves the commission of a crime within the jurisdiction of the Court; or
(ii) Be made in the knowledge of the intention of the group to commit the crime.127
Accordingly, common purpose liability lowers the requisite mens rea from purpose to knowledge, while preserving the same actus reus as aiding and abetting liability.
In addition to the mens rea of knowledge and the actus reus of material contribution, common purpose liability introduces the element of group criminality. The substantive offense to which liability attaches must have been committed by a group of persons acting with a common purpose. In other words, it must be a group crime. Because genocide and incitement to genocide are, by their nature, almost always committed by groups—specifically, state entities or media organizations—common purpose liability will almost always apply to these crimes.128 Therefore, in many cases, complicity in genocide or in incitement, under the Rome Statute, only requires that the defendant knowingly contributes to the substantive offense.
C. Complicity in Incitement
TOPIncitement’s status as a crime raises the question of whether it permits secondary liability. As scholars and commentators have acknowledged, incitement is often considered an “inchoate crime,” a punishable step toward the commission of another substantive offense (for example, the crime of attempted murder).129 As complicity does not generally attach to inchoate crimes, some contend a person cannot be complicit in inciting genocide.130 Indeed, the ICTR Trial Chamber suggested as much in Akayesu, noting in a footnote:
It appears from the travaux préparatoires of the Genocide Convention that only complicity in the completed offence of genocide was intended for punishment and not complicity in an attempt to commit genocide, complicity in incitement to commit genocide nor complicity in conspiracy to commit genocide, all of which were, in the eyes of some states, too vague to be punishable under the Convention.131
The ICTR Appeals Chamber, however, rejected this interpretation as to incitement, perhaps because any concerns around vagueness had been addressed by a budding jurisprudence. The Court provided that a defendant can be complicit in direct and public incitement to commit genocide, notwithstanding its character as an inchoate crime.132 The Appeals Chamber’s reasoning is consistent with the view that the effective prevention of genocide requires curbing its earliest stages.
In Nyiramasuhuko, the prosecution advanced the theory—before the ICTR Appeals Chamber—that defendant Joseph Kanyabashi had aided and abetted direct and public incitement to commit genocide.133 In that case, Prime Minister Kambanda and President Sindikubwabo delivered speeches which the prosecution argued to be incitements to genocide.134 Kanyabashi gave his own speech, in which he supported their message and pledged to execute the directives and instructions announced by Kambanda and Sindikubwabo.135 The ICTR Appeals Chamber ultimately rejected the prosecution’s argument, grounding its objection not in the impossibility of complicity in incitement but in its conclusion that defendant’s conduct did not meet the legal requirements for complicity in incitement.136
The ICTR Appeals Chamber noted incitement’s status as an inchoate crime, but nonetheless proceeded with a complicity in incitement analysis. Specifically, the Chamber held:
As an inchoate crime, direct and public incitement to commit genocide is completed as soon as the discourse is uttered or published, even though the effects of incitement may extend in time, and is punishable even if no act of genocide has resulted therefrom. Accordingly, in order for Kanyabashi to be found responsible for aiding and abetting direct and public incitement to commit genocide, it would have to be established that he substantially contributed to Kambanda’s and Sindikubwabo’s inciting speeches themselves and not, as the Prosecution suggests, to the effects of their incitements by “reiterat[ing] and reinforce[ing] their message.”137
The Chamber went on to note that because Kanyabashi’s speech occurred after the incitement, it could not have substantially contributed to its commission—suggesting that if the prosecution had shown that Kanyabashi substantially contributed to the inciting speeches, he could have been complicit in inciting genocide.138
Nyiramasuhuko therefore clarifies that an aider and abettor can be complicit in the direct and public incitement to commit genocide, notwithstanding incitement’s status as an inchoate crime. The extension of complicity to incitement is undoubtedly controversial, as it dramatically expands the scope of criminal conduct associated with a speech act. But there are two points that ought to curb this controversy to some extent.
First, genocide is the only substantive offense in all of ICL for which incitement is also criminalized.139 Article 25 of the Rome Statute makes this clear.140 While the knowing contributor to direct and public incitement to genocide may be implicated in an international crime, the same cannot be said for he who knowingly or purposefully contributes to the vast array of expressive conduct that falls short of incitement to genocide (for example, incitement to non-genocidal violence). This unique criminalization of complicity reflects the distinctive place that genocide holds in ICL. It also provides the ICC prosecutor with a powerful tool to punish those who knowingly contribute to inciting genocide, where the inciters are coordinating as a group.
Second, the ICC prosecutor has constrained resources and thus selects cases according to limiting principles, including the gravity of the crimes, the degree of responsibility of the alleged perpetrators, and the potential charges.141 Accordingly, consistent with the degree of culpability, those complicit in incitement may be less of a prosecutorial priority than those who incite genocide, who may be less of a priority than those who commit genocide. This may not be the case, however, with respect to social media companies, given their profound influence over the dissemination of information today.142
IV. Corporate Criminal Liability
TOPThus far, this Comment has discussed the most relevant international crimes (Section II) and the modes of liability most pertinent for later analysis (Section III). Given this Comment’s examination of whether social media companies risk international criminal liability when they grant inciters a platform, it is imperative to explore whether a corporation may be held responsible for international crimes and the implications of this inquiry on criminal liability.
Although scholars continue to debate the normative question of whether ICL should extend to corporations, ICL does not generally allow for their prosecution as collective entities.143 The Rome Statute states that “[t]he Court shall have jurisdiction over natural persons pursuant to this Statute.”144 The use of “natural persons” was no accident. The drafters of the Rome Statute disagreed about whether criminal liability ought to be extended to “legal persons,” knowing it would include corporations under its jurisdiction, but ultimately decided against this language.145
While corporations as collectives cannot be subjected to ICC prosecution, their individual employees can, even when engaging in business activity. According to Cassel, there is a long history of corporate executives being held criminally responsible under ICL.146 ICL has been applied in instances where a company’s employee committed a crime him or herself, as well as instances where that employee aided and abetted in the commission of a crime. Wolfgang Kaleck and Miriam Saage-Maaß endorse this view, writing that “[c]ase law shows that individuals within a corporation . . . can be held criminally liable for the commission” of an international crime “occurring in the process of ‘doing business.’”147 Such precedent dates back to the Nuremberg trials, which demonstrate that corporate executives are not immune from criminal prosecution. For example, in the principal Nuremberg case, “German industrialist Gustav Krupp was originally indicted along with top Nazi government, party and military leaders, and escaped prosecution only by reason of age and infirmity.”148 Prosecutors had stated that executives at German corporations were instrumental in supporting the Nazi party.
In The Zyklon B Case, the British Military Court in Hamburg examined whether a number of German businessmen were complicit in “the murder of interned allied civilians by means of poison gas.”149 The men faced trial for working at a company which supplied the Nazis with Zyklon B, a highly toxic gas used to murder Jews in concentration camps.150 Defendants included the firm’s owner Bruno Tesch, his “second-in-command” Karl Weinbacher, and gassing technician Joachim Drosihn.151 In their defense, Tesch and Weinbacher relied upon the fact that they were not present at the concentration camps.152 The prosecution contended this was immaterial, that knowingly supplying “a commodity to a branch of the State which was using that commodity for the mass extermination” of civilians was illegal.153 Citing reports from the firm’s own employees, prosecutors argued the defendants became aware that Zyklon B gas was being used for the “extermination of human beings” and that “having acquired this knowledge, they continued to arrange supplies of the gas” in “ever-increasing quantities.”154 Tesch and Weinbacher were found guilty, while Drosihn was acquitted.155 The Court concluded that Tesch and Weinbacher, as the company’s top executives, materially contributed to the crime when their business continued supplying gas to the Nazis and that they did so knowing the gas would be used to kill human beings.156 In contrast, Drosihn, a mere technician, lacked the authority “to influence the transfer of gas . . . or to prevent it.”157 He was therefore acquitted.158
Beyond the Rome Statute and the relevant precedent, the potential culpability of individuals operating through businesses traces back to earlier treaties. The Genocide Convention unequivocally states that “private individuals” may be punished for the commission of Article III acts, such as genocide and direct and public incitement to commit genocide.159 The ICTY and ICTR Statutes granted jurisdiction over “natural persons,” which necessarily include individuals acting on behalf of businesses.160
Thus, this Comment proceeds having established that individuals acting on behalf of corporations are fully capable of committing international crimes. As demonstrated by The Zyklon B Case, such individuals may also be complicit in the commission of a substantive offense.
V. How Social Media Companies Could Be Complicit in Incitement to Genocide
TOPThus far, this Comment has concluded the following: (1) the direct and public incitement to commit genocide is a crime; (2) an actor may be complicit in the direct and public incitement to commit genocide via aiding and abetting liability or, as is more likely, common purpose liability; and (3) while international criminal liability cannot extend to a corporation, liability can reach individuals working on behalf of that corporation. Applying these principles, this Section examines whether social media companies put their employees at risk of international criminal liability when they provide inciters with a platform. Specifically, this Comment illustrates—through the example of Facebook in Myanmar—how complicity would apply to individuals at three levels of the corporate hierarchy: the Chief Executive Officer (CEO), the manager, and the content moderator.161
By applying the theory of complicity in incitement to the foregoing cases, this Section highlights the sweeping ramifications of this novel theory, as well as its limitations. Such results ought to provoke broader discussion about whether the ICC should more seriously consider the prosecution of social media executives and how social media companies can reduce their employees’ exposure to criminal liability.
A. Inciting Genocide on Social Media
TOPAs a threshold matter, a social media company’s employee cannot be complicit in directly and publicly inciting genocide unless the incitement has been committed on social media. Four elements must be satisfied in order for someone to be guilty of incitement: (1) the incitement must be intentional (mens rea); 2) it must be public; (3) it must be direct; and (4) the defendant must have the specific intent to cause genocide.
The question of whether an incitement is intentional—distinct from the question of specific intent—turns on whether the statement itself is intentional. Accordingly, in the context of social media, an incitement will almost always be intentional, given that one rarely posts online accidentally.
Whether an incitement is public or direct is a more difficult inquiry. But because social media is, by its nature, a “technological means of mass communication,” proving incitement as public is less of an obstacle.162 In this way, incitements on social media are no different from Streicher’s weekly newspaper or Nahimana’s radio broadcast.163 They all involve the widespread dissemination of information to external audiences and are therefore all public. There are some notable distinctions, however, between social media and these other mediums, at least with respect to certain applications of social media. For example, using social media for private messaging would probably not be public. Moreover, if a social media user has tailored his or her privacy settings such that the incitement is viewable only among a small number of individuals, the post may not satisfy the public requirement.164 But generally speaking, the use of social media to post incitements for others to view will likely be public. With respect to Myanmar, a 2018 report by the U.N. Independent International Fact-Finding Mission on Myanmar (IIFFMM) revealed that Tatmadaw officials routinely used social media to disseminate information about the Rohingya to nationwide audiences.165 Any incitement was therefore public, more similar to a newspaper or a radio broadcast than private correspondence among a small number of individuals.
The next question is whether the incitement is direct. As stipulated in Section II.B., while a vague suggestion is not direct, an incitement need not be an explicit call for genocide.166 Where the incitement is not explicit, the court reviews cultural context to determine whether the audience clearly understood the statement as a call for genocide.167 To that effect, a systematic propaganda campaign to dehumanize a protected group may constitute direct incitement.168 Statements which in isolation fall short of incitement may, in the aggregate, satisfy the direct requirement. This approach is consistent with a multi-staged conception of genocide, which begins with the systematic dehumanization of a protected group. It recognizes that a campaign of implicit statements may more effectively incite genocide than one explicit call. The courts in Streicher and in the Media Case understood as much, emphasizing the systematic nature of the incitements in those cases.
Social media plainly allows for the possibility of direct incitements. Like print newspapers and radio broadcasts, social media provides inciters with a platform to disseminate their messages. Social media may actually enhance the inciter’s capacity to dehumanize a protected group. Unlike conventional media, it deploys algorithms with a tendency to amplify content tailored to individual predispositions, whatever they may be.169 Such algorithms may contribute to the exacerbation of existing social tensions within a nation.170 In extreme cases, this could result in the dehumanization of a protected group and therefore help set the social conditions necessary for genocide.171 But even if a court ignored the role of algorithms, social media’s capacity to circulate incitements is sufficiently analogous to the conventional media used in Streicher and the Media Case to facilitate direct incitement, through either the explicit call for genocide or the implicit but systematic campaign to dehumanize a protected group.172
In Myanmar, the IIFFMM found that Tatmadaw officials used Facebook to execute a systematic campaign to dehumanize the Rohingya.173 Officials actively promoted the narrative that the Rohingya did not exist in Myanmar. For example, on September 1, 2017, one senior official stated “so we openly declare that ‘absolutely, our country has no Rohingya race.’”174 Moreover, officials reinforced the “narrative of the whole Rohingya population being ‘terrorists’ and inherently violent.”175 The IIFFMM also found that “[m]ost of the Myanmar authorities’ posts and communications [ ] directly [fed] the narratives of illegal immigration and Islamic threat.”176 Even if these statements only implicitly dehumanized the Rohingya, the IIFFMM produced evidence that the audience potentially understood the posts as calls for genocide, citing user comments on Tatmadaw posts.177 While further analysis is required to discern whether the Tatmadaw’s campaign constituted direct incitement, an implicit but systematic effort to dehumanize the Rohingya may be sufficient under Streicher and the Media Case, where similar tactics resulted in direct incitement.178
Finally, a prosecutor must show the alleged inciter had the specific intent to cause genocide. Such intent may be inferred from the circumstances.179 Where a genocide actually occurs, this inquiry may be straightforward, as the genocide itself provides strong evidence of specific intent.180 Nonetheless, the content of the post or the systematic nature of the incitements may also suggest specific intent. International courts are reluctant to find specific intent where the individual merely passed along content developed by another, as evidenced by the acquittals in Fritzsche and the Media Case.181 Accordingly, individuals on social media who merely promote or share inciting content spontaneously, and not as part of some organized campaign, are unlikely to have the specific intent necessary for an incitement conviction. Applying these principles to Myanmar, it may be difficult to show that citizens who shared Tatmadaw Facebook posts, even where the post dehumanized the Rohingya, committed incitement. But if the ICC concludes that Tatmadaw officials committed genocide, the specific intent argument might be strengthened.
B. Complicity in Inciting Genocide on Social Media
TOPOnce direct and public incitement to commit genocide is established, the potential complicity of others becomes an open question, and incitements on social media may implicate employees at the company providing the platform. As discussed in Section III, there are two main tracks for complicity under the Rome Statute: (1) aiding and abetting liability; and (2) common purpose liability.182 Because the requisite mens rea for common purpose liability is knowledge, while the mens rea for aiding and abetting is purpose, a prosecutor is more likely to obtain a conviction where the incitement is perpetrated by a “group of persons acting with a common purpose.”183 Therefore, an employee faces a greater risk of complicity in an incitement campaign—which requires coordination among multiple actors—than an isolated call for genocide.184 This Section’s analysis is therefore limited to complicity in campaigns, for which the mens rea is knowledge. Accordingly, a social media employee who knowingly assists a campaign to incite genocide is complicit in that crime.185 As evidenced by The Zyklon B Case, complicity also varies depending on where an individual is positioned in the corporate hierarchy.186 Accordingly, this Comment examines the risk of complicity through three examples: the CEO, the manager, and the content moderator.
1. The CEO
This Comment first analyzes the CEO. It is presumed the CEO has the foremost control over the social media company, setting the overall direction for the enterprise and making fundamental business decisions. While the CEO’s awareness of business activity is likely broader than his or her subordinates with respect to scope, he or she is less likely to know the day-to-day details of the company.
First, a prosecutor must show that the social media CEO assisted the incitement. According to the ICC, assistance may be practical or material, as well as moral or psychological.187 Since this case involves the assistance of business executives, The Zyklon B Case serves as instructive precedent. The social media CEO is most analogous to Bruno Tesch, the owner of the firm in that case.188 There, the Court held that Tesch assisted the Nazis, reasoning that his company provided the Nazis with material means (toxic gas) for the crime’s commission (killing via gas chambers). Similarly, the CEO of a social media company satisfies the actus reus of complicity if he or she provides a platform, the material means, to individuals executing a systematic propaganda campaign which amounts to incitement, the crime. The extent to which the CEO is involved likely bears on the materiality of the contribution. For example, one could imagine a scenario where a CEO delegates authority to expand service to certain countries. This CEO’s contribution to incitement would be less material than the CEO who personally pushes to expand social media services. In the case of Myanmar, publicly available information suggests that Facebook CEO Mark Zuckerberg falls into the latter camp and may have satisfied the actus reus for complicity.
In August 2013, Zuckerberg personally announced an initiative called “Internet.org,” “a plan to make the internet available for the first time to billions of people in developing countries.”189 The CEO unveiled the proposal in a 10-page white paper, posted on Facebook.190 Zuckerberg framed the plan as an attempt to make “internet access available to those who cannot currently afford it.”191 Consistent with this initiative, Facebook launched a Myanmar version of the platform in 2015, although a version of the platform had been available years prior. The service quickly attracted millions of users.192 According to the IIFFMM, the “relative unfamiliarity of the population with the Internet and with digital platforms and the easier and cheaper access to Facebook led to a situation in Myanmar where Facebook is the Internet.”193 The IIFFMM further noted that the platform became “a regularly used tool for the Myanmar authorities to reach the public.”194
Several years before 2018, the year Facebook began taking down official accounts, the Tatmadaw launched a systematic campaign which involved “hundreds of military personnel who created troll accounts and news and celebrity pages on Facebook and then flooded them with incendiary comments and posts timed for peak viewership.”195 Tatmadaw officials repeatedly reinforced and legitimized dehumanizing themes and narratives with more opaque posts from their official accounts.196 Ultimately, the campaign culminated in the mass killings of the Rohingya.197 The foregoing reports, to the extent that they are accurate, suggest that by launching Internet.org and extending Facebook’s services to Myanmar, Zuckerberg provided the Tatmadaw with a massive platform to systematically promote dehumanizing narratives about the Rohingya. Like the CEO’s contribution in The Zyklon B Case, Zuckerberg extended a commercial service that proved indispensable to the commission of an international crime (assuming the Tatmadaw’s campaign amounted to incitement).198 Accordingly, it is at least plausible that Zuckerberg satisfied the actus reus for complicity.
The next pertinent inquiry is whether the CEO made the contribution knowingly. The CEO is only complicit if he or she knew the inciters were using the platform to commit direct and public incitement to genocide.199 Reflective of this challenge, much of the litigation in The Zyklon B Case centered around the issue of mens rea.200 The prosecution succeeded by pointing to multiple reports by individual employees, as well as the fact that the company’s gas shipments increased so significantly that Tesch must have known the end to which they would be used.201 But the size of social media companies alone would make it difficult for the CEO to know the particulars of any given business dealing.202 Presumably, there are everyday business operations of which the CEO has no knowledge. And flaws in the company’s compliance system, while arguably evidence of negligence or recklessness, could insulate the CEO from satisfying the higher mens rea of knowledge.
On the other hand, even if a CEO’s subordinates do not report the incitement to genocide, the CEO could be alerted to the situation by external sources. For example, a public official or the leader of a human rights group could have channels to directly notify the CEO. Such a notification would need to demonstrate the existence of a systematic campaign to incite genocide on the CEO’s platform. Isolated notices of hate speech or deeply offensive language would be insufficient. Assuming the veracity of such a report, the CEO’s subsequent actions would bear heavily on his or her culpability. Once he or she has the requisite mens rea of knowledge, continuing to provide inciters with a platform would merit a finding of complicity.
With respect to Myanmar, the operative question is whether Zuckerberg knew of the Tatmadaw’s campaign against the Rohingya. Whereas the case for actus reus seems plausible, the case for mens rea is less convincing, at least based on publicly available information. After Zuckerberg announced Internet.org, the initiative that eventually brought Facebook to Myanmar, he said the following in an interview with WIRED: “Our service is free, and there aren’t developed ad markets in a lot of these countries. So for a very long time this may not be profitable for us. But I’m willing to make that investment because I think it’s really good for the world.”203 Many of Zuckerberg’s public comments around the Internet.org announcement and thereafter reflect a similar optimism about the potential of bringing Facebook to Myanmar.204 Misguided as it may have been, such remarks do not evince knowledge that the Tatmadaw would use the platform to incite genocide. And according to The New York Times, the propaganda campaign went “undetected.”205 Moreover, the company’s scant devotion of translation resources to Myanmar suggests that Zuckerberg did not personally know about the ongoing and systematic attempt to dehumanize the Rohingya.206 Assuming these reports are true, Facebook’s CEO does not appear to satisfy the requisite mens rea. Thus, even if Zuckerberg satisfied the relevant actus reus, a finding of complicity would be unmerited, although further investigation would be necessary to draw a firm conclusion either way.
2. The manager
In contrast to the CEO, the manager does not direct all aspects of the social media company. For the sake of this analysis, he or she is placed in a senior-level position in the company, with a more defined portfolio of responsibilities. The manager directs a narrower set of operations and, as such, is more likely than the CEO to have expertise on certain aspects of the company. Furthermore, the manager may be vested with decision-making authority regarding controversial issues (for example, whether to remove a government official’s social media account). Whereas a content moderator may handle day-to-day review of material posted online, a manager could be responsible for synthesizing information and for detecting broader trends on the platform. This analysis also assumes that the manager personally spearheads certain business operations, like the accessibility of social media to particular countries.
With respect to assistance, the case may be clearest for the manager. Unlike the CEO, the manager would personally lead any effort to expand services into certain countries and, as such, may be vested with critical decision-making authority (for example, determining whether the costs associated with expansion outweigh the benefits). Drawing parallels to The Zyklon B Case, the manager could be analogized to Karl Weinbacher, the senior executive and manager convicted of complicity, because of the extensive personal involvement in operationalizing the harmful business initiative.207 Accordingly, the manager’s assistance to potential inciters may be less attenuated than the CEO’s, especially where the manager has the discretion to make a platform available in the first place. In satisfying the actus reus of assistance, however, it is important to note this analysis assumes a manager with a substantial degree of autonomy and a relatively deferential CEO. To the extent that the CEO supersedes the manager’s responsibilities, the case is weaker against the manager and stronger against the CEO.
Beyond the actus reus of assistance, the manager is relatively well-positioned to know when the company is providing a platform to inciters.208 Whereas a CEO’s broader responsibilities may insulate him or her from awareness of ongoing incitement, responsibility over the platform’s growth and sustainability in certain countries may fall squarely within the manager’s portfolio. Public reports of Facebook’s activities in Myanmar highlight this distinction. While it remains unclear whether Zuckerberg knew of ongoing incitement, a special report by Reuters revealed that senior Facebook officials were warned over a span of years “that [Facebook] was being used in Myanmar to promote racism and hatred of Muslims, in particular the Rohingya.”209 For example, a tech entrepreneur who worked in Myanmar “said he told Facebook officials in 2015 that its platform was being exploited to foment hatred in a talk he gave at its headquarters in Menlo Park, California.”210 Such warnings were reportedly made years before Facebook began taking down Tatmadaw accounts in 2018. The operative question would be whether, pending further investigation, these notices illuminated an effort by the Tatmadaw to incite genocide. If so, a prosecutor could potentially argue that by providing the Tatmadaw with a massive platform, with knowledge of the campaign, managers at Facebook satisfied the bar for complicity in inciting genocide. The mere possibility of this argument suggests that managers could face substantial risk of criminal liability.
While managers akin to senior executives risk criminal liability where the facts support a finding of complicity in incitement, the case is likely weaker for managers at lower tiers of the company. The foregoing analysis has used the term “manager” as a proxy for executives with authority similar to Weinbacher in The Zyklon B Case. Therefore, a mid-level supervisor at Facebook would not necessarily face substantial risk, as the actus reus and mens rea for incitement might be harder to satisfy for such an employee.
3. The content moderator
Having examined the CEO and the manager’s potential exposure to international criminal liability, this Comment examines whether a social media company’s content moderator could be complicit in inciting genocide. The moderator is a low-level employee responsible for reviewing user-generated content and making daily decisions as to whether certain content violates the company’s community standards. As social media companies have increasingly globalized their services, it has become more common for moderator responsibilities to include translation.211
With respect to assistance, the moderator helps maintain the health of the online community.212 While the failure of a moderator to remove an incitement or an inciter from the platform might indirectly assist perpetrators, the moderator does not decide whether certain countries, or individuals therein, will be granted initial access to the platform. This lack of material decision-making authority makes the moderator most comparable to Joachim Drosihn in The Zyklon B Case, the gassing technician who was acquitted.213 As the Court acknowledged in that case, low rank in the corporate hierarchy provides some degree of immunity with respect to complicity when the assistance rendered stems from business dealings involving more senior executives.
Assuming a prosecutor surmounts the high bar of proving a moderator sufficiently assisted an inciter, it would be an additional obstacle to show the moderator knew of a coordinated effort to incite genocide. Since the moderator is generally part of a broader team, each individually tasked with content review, the responsibility of detecting and addressing more threatening patterns, like a systematic campaign to incite genocide, likely requires more expertise and so would fall to a more senior employee. Finally, even if a prosecutor could prevail in showing assistance and knowledge, the deterrence effect of such a prosecution would be minimal. Accordingly, the moderator faces the least risk of complicity.
4. Normative considerations
In sum, the foregoing cases demonstrate how social media employees at varying levels of the corporate hierarchy could be complicit in inciting genocide. While low-level employees face little risk of criminal liability, manager-level employees and CEOs open themselves up to substantial risk when they knowingly provide a platform to the perpetrators of incitement. Since CEOs may be more insulated and less involved in the expansion of services than managers, they could face less risk depending on the particular facts. Whether or not these results are satisfying, they ought to provoke a broader discussion about whether the ICC should more seriously consider the prosecution of social media executives and how social media companies can reduce their employees’ exposure to criminal liability. These questions are challenging, but they cannot be ignored, not after what happened in Myanmar. Cognizant of this complexity, this Comment offers three recommendations to help anchor future discussion.
First, the Rome Statute’s limitation on “natural persons,” listed in Article 25(1), should be amended to “legal persons,” thereby extending criminal liability to corporations. Such an amendment would enable the ICC prosecutor to focus on social media companies as distinct entities, as opposed to individuals whose prosecution may do little to change corporate behavior.
Second, the U.N.—together with the ICC—should work to produce a set of regulatory guidelines for social media companies that make their platforms available to new countries. Such guidelines would clarify the potential risk of international criminal liability posed by such ventures. These guidelines must be narrowly tailored to the crime of direct and public incitement to commit genocide, so as not to incentivize the over-policing by social media companies of speech acts which fall short of incitement.
Third, social media companies should invest more heavily in efforts to identify disinformation campaigns and bolster their content removal capabilities, so as to mitigate the potentially deadly effects of propagation. In order to determine whether content constitutes direct and public incitement to commit genocide, social media companies should ensure they have content moderators who not only understand the language of a country in which their platform is available but also sufficiently understand cultural context to know when implicit statements may be indicative of calls for genocide. Social media companies should also continue building partnerships with civil society organizations that can serve as an additional source of early alerts that an incitement campaign is underway. Such partnerships must not be considered replacements, however, for internal mechanisms to identify and root out incitement.
This Comment examined whether social media companies and their employees risk international criminal liability when they provide a platform to the perpetrators of direct and public incitement to commit genocide. In doing so, this Comment took a substantive approach rather than a procedural one, which might examine issues such as jurisdiction. It explored the ICL around genocide and the incitement to genocide (Section II), complicity (Section III), and corporate liability (Section IV) to conclude that although a social media company cannot be implicated in a crime as a distinct entity, individual employees at these companies can be complicit in inciting genocide. This Comment took the position, citing precedent, that complicity in incitement is not only a valid legal theory under ICL but also a potentially powerful tool for the ICC prosecutor to combat genocide’s early stages. This Comment also considered how complicity in incitement applies at three levels in the hierarchy of a social media company, including the CEO, the manager, and the content moderator (Section V). This analysis ought to provoke broader discussion about how ICL might be reformed to deter future incitements to genocide. Accordingly, this Comment concluded with a brief list of recommendations to serve as a starting point for future dialogue.