In April 2021, AI Dungeon implemented a new algorithm for content moderation to prevent instances of text-based simulated child pornography created by users. The moderation process involved a human moderator reading through private stories.[49][41][50][51] The filter frequently flagged false positives due to wording (terms like "eight-year-old laptop" misinterpreted as the age of a child), affecting both pornographic and non-pornographic stories. Controversy and review bombing of AI Dungeon occurred as a result of the moderation system, citing false positives and a lack of communication between Latitude and its user base following the change.[40]
Haha. Good find.