How does some random webshit showing stuff possibly relevant to a user constitute being a radicalization engine? Is this something they could have known at design time? I'm not being rude, I want to know the reasoning.
In this case, Meta was most definitely complicit for other reasons.
- they started operating in a country but had no Myanmar-language human moderation, in a place where FB quickly became synonymous THE internet.
- the AI-powered content moderation failed hard on identifying hate speech partially because everything was written in Zawgyi, not Unicode.
facebook also doesn't have any content moderators in navajo or maya. should they block indigenous people from using their native languages on facebook to avoid harming them?
Interesting question.... but it is a very different case than voluntary starting to operate in a country with voluntary no investment in an ok moderation in the main language spoken there
- they started operating in a country but had no Myanmar-language human moderation, in a place where FB quickly became synonymous THE internet. - the AI-powered content moderation failed hard on identifying hate speech partially because everything was written in Zawgyi, not Unicode.
It was way after the FB had a lot of eggs on their face that they issued public statements: https://about.fb.com/news/2018/11/myanmar-hria/
They should have been much more proactive, but it seems like it was growth at all costs and ignore the rest.