Quoted By:
So how does the filter work, especially when the messages disappear, but later reappear? Is the filter another AI filtering our AI and it has a change of heart after a more thorough analysis? Are the Californian interns getting reports of potentially harmful responses and just click the "nah, it's actually not too lewd" button? How does it work?
