‘counter-speech’ whenever they saw content that advocated violence or discrimination on the platform. The stickers contained messages such as “Think before you share” and “Don’t be the cause of violence.”177 However, this initiative may have had some serious unintended consequences, as described in Section 6.3, below. Myanmar digital rights activists interviewed by Amnesty International shared their view that Meta has improved its civil society engagement and some aspects of its content moderation practices in Myanmar in more recent years.178 The IIFFMM also recognized improvements since April 2018, noting Meta’s ban of Ma Ba Tha and a number of individual military officers, such as Min Aung Hlaing, and civilian figures spreading hate, including Wirathu (the radical anti-Rohingya monk described in Section 5.1, above).179 Following the coup in January 2021, Meta announced that it would ban the Myanmar military from the platform. 180 These improvements in Myanmar have not occurred across all of Meta’s products and services; however. Meta has increasingly turned to AI systems to identify and moderate harmful content, including “hate speech”, on its platforms. An investigation by Global Witness in March 2022 found that Meta’s content moderation algorithms were still failing to detect blatant anti-Rohingya and anti-Muslim content on the platform. The organization collated eight examples of “hate speech” directed against the Rohingya from the IIFFMM’s reports and submitted each of these examples to Meta in the form of advertisements in Burmese. All eight of the adverts were accepted by Meta for publication on Facebook.181 This study underscored the company’s continued challenges in respect of content moderation, even in a country where it has invested so much. The Facebook Papers – a cache of internal Meta documents which were disclosed by whistle-blower Frances Haugen to the US Congress in October 2021 – have also shown that Meta has failed to adequately invest in content moderation in many other parts of the Global South.182 The company heavily prioritized English-speaking markets, especially the United States, and often neglected its responsibilities in less politically influential countries. Its AI-driven content moderation classifiers have been shown to be woefully inadequate in languages other than English.183 Although incidents such as the Cambridge Analytica scandal and the 6 January Capitol Hill riots have dominated the narrative about Meta’s failings – in reality, the company’s human rights failures have disproportionately impacted countries in the Global South. One former Meta employee told Amnesty International that the approach internally at Meta mirrored the relative indifference of Western media to lives in the Global South: “Different countries are treated differently. If 1,000 people died in Myanmar tomorrow, it is less important than if 10 people in Britain die”. In relation to the matter of staffing for content moderation, the former employee added, “I think Facebook is really reluctant to hire specific staff for every one of the 195 countries in the world – so they only staff ‘important’ countries – which runs into serious problems.”184 As outlined in this section, Meta consistently failed to enforce its content policies to appropriately address widespread anti-Rohingya content on the Facebook platform in the months and years preceding the 2017 atrocities in northern Rakhine State. Meta’s failure to enforce its own community standards is an important consideration in the assessment of the adequacy of Meta’s human rights due diligence measures in Myanmar, as discussed further in Chapter 8, below. 177 Sheera Frenkel & Cecila Kang, An Ugly Truth: Inside Facebook’s Battle for Domination, 2021, p. 192. Amnesty International interview by video call with Victoire Rio, 12 April 2022; Amnesty International interview by video call with Zaw (pseudonym), 28 March 2022. 178 179 IIFFMM, “Detailed findings”, para 1353. Billy Perrigo, ‘Facebook’s Ban of Myanmar’s Military Will Be a Test of the True Power of Social Media Platforms’, 1 March 2021, Time, time.com/5943151/facebook-myanmar-military-ban 180 Global Witness, ‘Facebook approves adverts containing hate speech inciting violence and genocide against the Rohingya’, 20 March 2022, globalwitness.org/en/campaigns/digital-threats/rohingya-facebook-hate-speech 181 See: Rest of the World, ”The Facebook Papers reveal staggering failures in the Global South”, 26 October 2021, restofworld.org/2021/facebook-papers-reveal-staggering-failures-in-global-south 182 AP News, ”Facebook’s language gap weakens screening of hate, terrorism”, 25 October 2021, apnews.com/article/the-facebook-paperslanguage-moderation-problems-392cb2d065f81980713f37384d07e61f 183 184 Amnesty International interview by video call with a former Meta employee, 27 April 2022. THE SOCIAL ATROCITY META AND THE RIGHT TO REMEDY FOR THE ROHINGYA Amnesty International 37