be left alive”. Facebook´s algorithms translation was then: “I shouldn´t have a rainbow in Myanmar” (Stecklow, 2018b). As a result of this issue, the translation-function was removed until further notice on the 28. August 2018, after just a short period of time (Stecklow, 2018a). Mimi and Pemala once told me how English translation was a problem. I had asked them about the Facebook posts some people write in English and why. Laughing, they looked at me and said that it is because of “likes,” they get more likes if they write in English. When asking if all of their Facebook friends were able to read it, referring to the level of English skills most people inhabit, Mimi answered: “Nooo, not all of friends.” Pemala continued: “They like it still, translation is bad, so they cannot understand all posts.” I did not think much of it at the time, but it turned out that this is part of a more significant issue. If the translations were accurate, it could have been, in addition to understanding posts in general, an essential tool in the fight against hate-speech. Proper translation could allow non-Burmese-speaking residents to report potential hateful content. The Facebook-system, the algorithms that usually works with constraining hate speech is not fully functional with Burmese script, and this is partly because of the font issues. This means that Facebook is more dependent than in most of the other countries they work in on their users to report content that is discriminating or violent in the platform, and that they depend on having enough employees to handle the reported content. In most parts of the world, this is functional, and together with the algorithms, it is sufficient. Why not in Myanmar? We already know that Facebook was too late to the game with sufficient Burmese-speaking employees, but what about the first step? For the employees to be able to do their job they rely on the fact that people actually report content to review. To then be able to function appropriately, Facebook relies on what Miller addresses as participatory surveillance.(Daniel Miller, 2011, p. 173). Alternatively, as Lambert puts it: “Facebook is a kind of ‘participatory panopticon’ in which users willingly submit to the ‘policing and establishing of normative behavior.’” (Lambert, 2013, p. 41) This allows the users, to some extent, surveil their environment in the feed. Through interactions and reporting content, they are part of the security system as well as establishing the social frames, contexts, and norms of the accepted behavior, or in this case – the sound of the content. 81