what really concerns me about this kind of automated algorithmic moderation is that it's not explicitly transparent what it means and how it reaches a decision to restrict certain content. if it's not open then how can I know I can trust it?

it has to be reported and then manually applied by the moderation team