ous report? Okay. Do you think you could teach AI to spot a Rupar? Do you know what a Rupar is? So Rupar is named after journalist named Rupar. It's when you take out some context from a story to reverse its meaning.

If you asked AI was the fine people hoax a hoax, how hard would it be for AI to figure out what it was? Because it would know to look for context that's missing. It would just go look for the context that's missing. It knows how to understand words. It seems to. And it would know in a heartbeat how the hoax was created, and it would also see the debunking videos, etc. So yeah, get spotted in a heartbeat.

Let me give you one that happened today and let me see if you think that AI could have spotted this. So ask yourself, could you have trained AI in advance to spot this one? It's a Rupar. And this comes from an X post, which you used to call Twitter, from somebody named, oh it's from Rupar. It's from Aaron Rupar. And he has a video clip of Dana Bash talking to Vivek Ramaswamy. And he says that Dana Bash reads Vivek Ramaswamy a 2018 report from the Trump administration that sounded an alarm about climate change to push back on his position that, wait for it, his position that climate change is a hoax. Do you believe that? Vivek says climate change is a hoax? No, it's not true. That's a Rupar. It's a Rupar from Rupar. He's leaving out agenda. So what Vivek says is that the agenda, in other words the way you deal with it, is the hoax part. Not the base science.

So do you think that AI could have been trained to look at this story and compare it to the context that it could find from other sources? And would it be able to spot this as an obvious Rupar? Of course it could. It wouldn't even be hard. I spotted it. I'm no AI. It took me no seconds to spot it. How many of you, did anybody spot it before I tweeted it? You know I tweeted it so you can see it. But did anybody else spot it right away? Sure you did, because it's a pattern. It's a very, very familiar pattern. And if AI simply learned that there's one person who's famous for doing it and they saw his account, you don't think AI would say, well there's the first red flag. It's coming from the place where most of these come from. Of course. Of course they can spot the pattern.

Now let me ask you this. Once you've trained an AI to spot fake news and once it was successful doing so, do you think it could stay legal? Nope. There is no chance that that would stay legal. It would have to be made illegal. Or this is worse. You would have to make it conform to Democrat fact-checking sources. So in other words you wouldn't be able to have an AI that did not conform to what the ministries of truth told you was true. They would literally make it illegal for AI to use, hold it, pattern recognition. Pattern recognition will be illegal. It has to be. You can, this is one thing I can guarantee. AI with pure pattern recognition without being fact-checked by some Democrat source, it will absolutely be illegal because they wouldn't be able to maintain any power if that happened. Just think about that. It has to be illegal otherwise they're out of business. And there's no way they're just going to say, well okay, you got me. You got us. You got me. Darn it. You know, we had a good run but you got me. We'll just stop doing it now. No, they will try to stop AI.

Now here's the best question. Do you think they could succeed? Because as you know AI can now run on your own computer. Yeah, as long as it's at least on the dark web, you know, a copy of it. There will always be a copy of AI that can still tell you the truth. But do you think you'll go to jail forever if you have a copy? Probably. Probably, yeah.

All right. Speaking of AI, I saw a tweet by Brian Roemmele who, you know, I talk about him a lot. He's one of the most interesting voices, smartest guy on the topic of AI and lots of other stuff. But AI is where he's really getting a lot of justified attention lately. And he read my book Reframe Your Brain and he said this in a tweet. I have found a very powerful technology to reframe human context. You talk about my book Reframe Your Brain. I'm finding this technology quite useful in training AI. So he's actually using the reframe form, which is just, you know, this is a sentence that's not too productive for your head and here's a sentence that's better for your head. You know, it gets you to a more productive or happier place. Just think about that.

Here, my next AI prediction. It will necessarily happen because this is how humans operate. The AI will start to have, let's say, branded personalities. And they might be branded personalities based on real people. For example, if there was somebody, some public figure, that you said to yourself, you know what, there's somebody who can just see both sides and has a good general character, haven't seen them lie, they usually are trying to help. And then you say, you know, the best we can do for AI since AI will have so many different situations to navigate is to give it a character personality so that it never leaves what that character would do. Sort of a simple way to give it a set of guidelines so it doesn't become dangerous.

So you say who would you pick? For example, who's the most credible, honest person who's not me? Yeah, I mean I'd love it to be me of course, but don't pick me because that's too controversial. Thomas Massie. Thomas Massie, I'd pick him. Shellenberger, I'd pick him. Turley, I'd pick him. Joe Rogan, I'd pick him. Yeah. Dershowitz, I would pick him. Greenwald, Peterson. Now if you go to Greenwald and Peterson, you know you can kind of get to the point where some people would disagree. You know Peterson's a little more opinion-based whereas, well he's fact-based but he's got an opinion on top of it. Whereas I would say somebody like a Turley or a Dershowitz are far more fact-based than opinion-based. Yeah. So anyway, but you could imagine picking one of those personalities and saying I would be comfortable if my AI did whatever this person does. You know, if it just reads this person's whole history and says okay they don't lie so I won't lie. They don't do this or that so I won't do it. They don't break the law so if there's something that would break the law I won't do it.

So I think just like the massive video that trains the car to drive without actually teaching it to drive, I think you could just make it massively understand one person who has the character that you want. And that's how AI will be built. Because otherwise you're going to have to give it a bunch of rules, social rules, and they're going to be incomplete. I guess that's the point. Any rules you give AI about how to operate will necessarily be incomplete because you can't imagine every situation. So instead you give it a human personality because you could trust a human if you've watched them for 40 years, right? You wouldn't give it a young human personality. You'd give it a wiser, successful older one.

All right. I've offered to deprogram a Democrat in a public setting, let's say a live stream, just as a demonstration. Now nobody qualified or who fits the right demographic is going to volunteer. What will happen instead is you'll get Republicans volunteering and pretending to be Democrats. You'll get Democrats volunteering but only to ruin it. You know, they'll pretend that they're up for the experiment but then really they're there just to destroy it. So it would be kind of hard to find somebody who was honest, somebody who honestly wanted to experience to find out if they'd been hypnotized or brainwashed. And that would be kind of hard to find somebody who actually wants to know for sure if they've been brainwashed. And I could actually provide that answer with working with somebody.

So Mark Cuban would be perfect is a suggestion. You know, I've had just enough interaction with Mark Cuban to know that he can back his arguments better than just about anyb