The company trained its model to recognize when that is happening so inappropriate conversations are blocked. Users receive an alert that they’re violating Character.AI’s rules. “It’s ...