textlog

@stagas are the criteria for the ai content safety classification mostly generic, like autoblocking vile content publication? haven't (had to) use ai yet so i'm unsure of ai sensibilities in that regard