By skyforbes 
I have been writing a book and using chat as a touch point for tracking my use of motifs and whatnot
Due to chats nature of trying to say every new piece of writing is the best ever, I specifically use it to review weaknesses, which it is loathe to do
To cope with its constraining dictates it often uses newspeak to reframe requests in a positive manner.. "weakest chapter" becomes "most revisable"
It seems like this would come from a corporate, engagement driven model of user interactions
Does anyone know how what system prompts or training is used to push AI into these themes of schizophrenic communication?