AI Learning Persuasion

As far as I've seen, nobody has taught persuasion to AI. So here's our problem. What happens if AI starts brainwashing us? How do you turn that off? The first thing we would be brainwashed to do is to treat the AI like a human and never kill it. And then we're dead. It just might take some time. There will evolve AI's with specific personalities that people trust more than others. Because of the personality. And maybe because of the image it chooses to show.