Alleged internal documents from Facebook appear to reveal “hate speech” guidelines which use formulas to determine what posts are deemed to be offensive content.
German publication SZ-Magazin posted internal documents reportedly from the Facebook department in charge of monitoring and deleting offensive content. The German paper states, “These rules are then often applied by external service providers, such as the Bertelsmann subsidiary Arvato in Berlin. Employees attend internal training workshops where they learn what should be deleted and what shouldn’t.”
The documents reveal that Facebook uses something called “protected categories” to define what counts as “hate speech” on the Facebook platform. The protected categories include sex, religion, nationality, ethnicity, race, sexual orientation, gender identity, and disability. It seems that any posts that discuss any of these categories in a negative manner can be classified as hate speech and deleted.
Age, employment status, continent of origin, social status, appearance, political affiliation, and religions are sorted into sub-categories that are given more room for error. These categories are generally allowed to be spoken about in a negative manner as long as a specific individual is not targeted over them.
Two protected categories when mentioned in the same sentence are sorted into a new category, the example given in the leaked documents states, “PC + PC = PC while PC + NPC = NPC”; for example Irish women = PC, but Irish teens = NPC.
The definitions relating to the discussion of refugees is also quite complicated. Religion is still enforced as a protected class, so writing something such as, “F*cking Muslims” is prohibited, but writing, “F*cking Migrants,” is not, as the term “migrants” is not part of a protected class.
The German publication further claims that as many as six hundred people in Berlin are working for minimum wage policing Facebook posts and comments and are apparently ill prepared for the workload they undertake on a daily basis. Some employees have even supposedly reported psychological issues due to some of the explicit imagery and videos they have had to censor from the platform.