or, you know, manipulate the results so that the pro-ai/ceo candidate is shown to be more appealing no matter what the actual situation of the conversation or facts are.

“oh no, sorry I was hallucinating”