“Given the potential harms, Facebook conducting business as usual in such a hot spot amounts to reckless social experimentation on a national scale by a USbased technology company, which had neither expertise nor physical presence in Myanmar. A baseline assessment would have determined that Facebook’s company mantra, “move fast and break things,” could exacerbate human rights risks, and that both its AI products and business strategy should be re-examined for the Myanmar market. If a company determined nonetheless to move forward in a human rights hot spot, subsequent reports confirming the risks would be sufficient evidence to support pausing and redirecting the company’s approach on human rights grounds.”36 b. Failure to account for actual adverse impacts, and to prevent and mitigate additional potential adverse impacts Not only did Facebook fail to conduct a baseline assessment to identify potential adverse human rights impacts of its services, it failed to prevent and mitigate additional adverse human rights impacts and account for actual ones as soon as they were brought to its attention. Facebook’s platform had previously been used to incite violence against the Rohingya and Muslims in the years leading up to the operations in 2012, 2014, and 2016. The IIFFMM points to widely disseminated posts from early in 2012 alleging that a group of Rohingya men murdered a Buddhist woman to have contributed to violence in Rakhine state, as well as similar posts alleging the rape of a Buddhist women by Muslims to have led to riots in 2014.37 Facebook was used to disseminate hate speech in the context of the 2016 operations, which resulted in the destruction of at least 1,500 Rohingya homes, as well as the forced displacement of more than 69,000 Rohingya into Bangladesh.38 According to The Guardian, digital researcher and analyst Raymond Serrato examined about 15,000 Facebook posts from supporters of the hard-line nationalist Ma Ba Tha group. The earliest posts dated from June 2016 (before the October 2016 operations) and spiked on 24 and 25 August 2017 when the operations happened. The analysis showed that activity within the anti-Rohingya group, which had 55,000 members, “exploded [during this time period,] with posts registering a 200% increase in interactions.”39 During these years preceding the clearance operations, activists and Burmese civil society warned Facebook about how its platform was being used to disseminate hate speech and disinformation against the Rohingya. From 2013-2015, activists and film-makers met with Facebook officials on this issue, even speaking directly to Facebook’s vice president of communications and public policy in 2014, and delivering a talk at Facebook’s headquarters 36 Ibid, p 6. IIFFMM, para 744, para 1347. 38 Human Rights Watch, ‘Burma: Security forces raped Rohingya women, girls’, 6 February 2017, <https://www.hrw.org/news/2017/02/06/burma-security-forces-raped-rohingya-women-girls>. [accessed 28 May 2021]; Andrew Selth, ‘Myanmar’s armed forces and the Rohingya crisis’, 17 August 2018, <https://www.usip.org/publications/2018/08/myanmars-armed-forces-and-rohingya-crisis>. [accessed 15 May 2021]. 39 Raymond Serrato, in Libby Hogan and Michael Safi, ‘Revealed: Facebook hate speech exploded in Myanmar during Rohingya crisis’, The Guardian, 2 April 2018, <https://www.theguardian.com/world/2018/apr/03/revealed-facebook-hate-speech-exploded-in-myanmarduring-rohingya-crisis>. [accessed 15 May 2021] 37 10