within ten minutes, but if they measure a larger group of people in a study, a lot of them will change their mind based on the AI trying to convince them what's true and what isn't.

Now, why is that? Hey, look who visited. This is Gary the cat. He will be joining us on the show today. If you'd like to look at Gary, I'll tip the camera down so you can look at the cat instead of me. Sort of an upgrade. Come on, get in here. All right, that's better.

So why do you think it would be true that AI would be persuasive? Let's see if I have taught you enough that you would have known that without this study. Number one, it's the documentary effect. If it's just you and the AI and the AI is one point of view, which typically it would, and it's trying to convince you that one point of view is right, and you spend ten minutes with it, there's a good chance it'll change your mind. Number two, there are no egos involved, or less of one. If you're dealing with another human, you're feeling like you don't want to seem less than them or dumb compared to them. So you don't want to change your mind because a human talked to you for ten minutes. That just wouldn't feel comfortable. But if you felt like the AI wasn't a person with an ego and it wasn't going to hold it over you if it was right and you were wrong, you'd never hear about it again. Well, then you would feel like you're just doing your own research and changing your mind on your own. So if you can get people to think that they're changing their mind on their own, it'll happen a lot easier than if it's like one human versus another human because you put your shields up in those cases.

So yes, it does make sense that AI would be super persuasive. Now, here's the troubling part. That super persuasion I believe happens without it knowing how to do persuasion. It happens just because it has a good argument and has good facts and people tend to believe the computer. So that's all it has. And it can already with no real persuasion technique, you know, just presenting arguments basically, it can already totally change people's opinions. What is going to happen when it starts using the techniques of persuasion? Because it knows them because it got trained on all the bodies of work in the world. So yeah, it knows what to do. But presumably it is not programmed to maximize persuasion, but it wouldn't take much to do it. So t