The connection between anti-Rohingya content online and the enactment of offline violence is an important part of any assessment of Meta’s responsibility for human rights abuses suffered by the Rohingya. As has been demonstrated, the mass dissemination of messages that advocated hatred inciting violence and discrimination against the Rohingya, as well as other dehumanizing and discriminatory anti-Rohingya content, was instrumental in increasing the likelihood of mass violence against the target group, with severe consequences for a wide range of human rights. This assessment is explored in detail in Chapter 8, below. 5.4 CONTENT MODERATION AND ANTI-ROHINGYA CONTENT ON FACEBOOK “Content moderation” refers to social media platforms’ oversight and enforcement of platform rules in relation to permissible and prohibited forms of expression. For Meta, these rules are known as “Community Standards”. This section provides an overview and analysis of Meta’s efforts to enact effective content moderation in the context of rising levels of anti-Rohingya content on the Facebook platform before and during the 2017 atrocities in Northern Rakhine State. Content moderation typically involves the detection of harmful content (such as “hate speech”) and then a range of varied actions in response, such as removal of the content, making the content less visible (‘demoting’ or ‘downranking’), or penalties against individual users or groups. Content moderation is enacted via a combination of human content moderators and content moderation algorithms.161 Meta has long had internal policies against the use of “hate speech” on its platform. The company’s community standards recognize the fact that hate speech on the platform can lead to offline violence and harms, stating: “…[W]e don’t allow hate speech on Facebook. It creates an environment of intimidation and exclusion, and in some cases may promote offline violence”.162 In 2018, as Meta faced increasing scrutiny for its role in the commission of crimes under international law against the Rohingya, Mark Zuckerberg was questioned in a hearing at the US Senate about the company’s role in the possible genocide against the Rohingya. He responded: “Senator, what's happening in Myanmar is a terrible tragedy, and we need to do more ... we’re hiring dozens of more Burmese-language content reviewers, because hate speech is very language-specific. It's hard to do it without people who speak the local language, and we need to ramp up our effort there dramatically.”163 Meta’s failure to invest in adequate content moderation in Myanmar has been heavily criticized as new details have come to light. Meta has never disclosed the precise number of Burmese-language content moderators it employed during the 2017 atrocities, but in a written response to US legislators in June 2018, Meta claimed to have “added dozens more Burmese language reviewers” over the prior two years.164 However, according to one investigation, Meta had only five Burmese language speakers to monitor and moderate content for Myanmar’s 18 million Facebook users at the time of Zuckerberg’s testimony to the US Senate in April 2018, with none of these individuals based in Myanmar.165 Even this small number of staff marked a significant improvement on the situation as it stood prior to the 2017 crisis. In mid-2014, Meta staff admitted in a secret Facebook group containing a number of Myanmar civil society activists that they only had one single Burmese-speaking content moderator devoted to Myanmar at the time, based in their For a detailed overview of content moderation from a human rights perspective, see: Ranking Digital Rights, ‘It's Not Just the Content, It's the Business Model: Democracy’s Online Speech Challenge’, 17 Mach 2020, newamerica.org/oti/reports/its-not-just-content-its-businessmodel 161 Facebook’s definition of “hate speech” differs from the definition of “advocacy of hatred” under Article 20 of the ICCPR, with Facebook’s definition encompassing a broader range of expression that would not always reach the threshold in Article 20. In June 2017, Facebook stated that its definition of hate speech at the time was “anything that directly attacks people based on what are known as their ‘protected characteristics’” and affirmed that it did not allow hate speech on the platform. See Meta, Hard Questions: Who Should Decide What Is Hate Speech in an Online Global Community?, 27 June 2017, about.fb.com/news/2017/06/hard-questions-hate-speech 162 The Washington Post, “Transcript of Mark Zuckerberg’s Senate Hearing”, 10 April 2018, washingtonpost.com/news/theswitch/wp/2018/04/10/transcript-of-mark-zuckerbergs-senate-hearing 163 Facebook Inc., Written Responses to Post-Hearing Questions by the United States Senate Committee on Commerce, Science, and Transportation in relation to the Senate Hearing on Facebook, Social Media Privacy, and the Use and Abuse of Data, p.136, available at: https://www.judiciary.senate.gov/imo/media/doc/Zuckerberg%20Responses%20to%20Commerce%20Committee%20QFRs1.pdf 165 Cecilia Kang and Sheera Frenkel, An Ugly Truth: Inside Facebook's Battle for Domination, Chapter 9, p.191. 164 THE SOCIAL ATROCITY META AND THE RIGHT TO REMEDY FOR THE ROHINGYA Amnesty International 34