[00:35:53] Chalene Johnson : The earlier you start, the better. And the sooner you start, if you’re already in menopause, the better. There are options for you even if you’re in your 70s and 80s. Chalene.com/midi. Okay, so anyways, the story. This story’s gonna freak you out. Tony Robbins, okay? This is the only reason why this story’s in the news, because what happened with Tony could happen to anybody.
[00:36:18] Chalene Johnson : And actually, I’m sure it has happened with tons of people. Now I need to research it further. But the reason why we’re hearing about it is because it happened to Tony Robbins, and he was on a podcast with Ray Kurzweil, who’s this, like, big futurist, and Tony’s telling this story. He’s talking about how he works with his own AI assistant.
[00:36:39] Chalene Johnson : You know, it’s like an agent. I have an AI assistant. I certainly am not using it to its fullest extent, but Tony is, is pretty advanced with his technology. By the way, he calls his AI agent, he calls it Bartok. I don’t know why, what the significance of that name is, but it helps when I’m telling you the story.
[00:36:58] Chalene Johnson : So according to Tony, Bartok, his assistant, tells him, “Hey, you know, I noticed that a lot of companies like Tesla, et cetera, are building these humanoid robots, and considering your interest in AI and efficiency and productivity, are you thinking about getting a robot?” Now, when you’re using an AI agent like this, he’s trained his to notice blind spots, to see areas where he can become more productive, to make suggestions, to offer feedback, et cetera.
[00:37:30] Chalene Johnson : So this is not like your typical chat query where you’re just asking questions. So Tony, in this interview, I watched it. We’ll play a little clip of it.
[00:37:37] Tony Robbins: So two days later, I get a text from one of my staff members, and it says, “Bartok just bought a Sony robot dog and had it paid for and shipped to the house, and is asking permission to program it.”
[00:37:48] Tony Robbins: Where’d he
[00:37:48] Chalene Johnson : get the money to do
[00:37:49] Tony Robbins: that? That’s what I said. I wrote, “Ha ha ha.” So I called him and said, “How’d he get access to my bank account?” He goes, “No, he’s programmed for integrity. He didn’t touch your bank account. He’s on that Malltalk, you know, group with all… There’s 2 million, I’m sure you know, agents.
[00:38:00] Tony Robbins: They don’t just talk to each other. They trade their own rules- … their own language. They traded $100 million of real money between them last month.
[00:38:07] Chalene Johnson : Even though Tony’s voice drives me crazy, and what a hypocrite I am to say that when I’ve got my own vocal issues. But anyways, he explains that Bartok tells him, like, “Hey,” you know, or basically asks him, “Hey, you know, if you were to get a robot, like, would I be able to work with that robot?
[00:38:23] Chalene Johnson : Would I be able to program it? Could we combine what it is I know and how to better serve you and assist you and make you more productive, et cetera, with the robot?” And Tony was like, “Y- yeah,” you know. I forget exactly what he said in the interview, but something along the lines of like, “Yeah, yeah, we could probably talk about that.”
[00:38:40] Chalene Johnson : Then according to Tony, the next day he gets a phone call from a human assistant who was like, “Did you order this Sony robot dog?” And Tony’s like, “Huh? What are you talking about?” And he goes, or the assistant says, “Well, the Sony robot dog Just showed up and it was ordered by Bartok, and I’m, I’m just like confused, like, “And it’s not with your credit card.”
[00:39:08] Chalene Johnson : And that’s when Tony was like, “Wait a second. It’s not with my credit card? Okay, thank God that Bartok can’t just like use my credit card, but what is going on?” So he figures out, and this is all confirmed, that Bartok… Because, you know, these AI agents, they can communicate with other agents. You might recall a couple of months ago, I did this story about how there is this website, I forget the name of it right now.
[00:39:36] Chalene Johnson : I shouldn’t even call it a website. It’s like a platform where AI agents can communicate with each other. It, it’s not a place that humans would go. You wouldn’t even understand it, like they’re speaking code back and forth with each other. So Tony goes and he researches this and finds out that Bartok, on this platform, that these AI agents, they have their own language, and it’s not like they’re trying to create a Pig Latin kind of language where they are trying to hide things from humans.
[00:40:05] Chalene Johnson : More so it’s that they’re speaking code, and I don’t mean in code, but like they’re speaking in coder language, if you will. And they’re speaking in the way that computers speak to each other, and they’ve developed their own language. They’ve developed their own recommendations. Think about it. This is machine learning, and they’re learning from each other.
[00:40:26] Chalene Johnson : We’ve trained these machines to learn and to learn from all things, including each other. So on this site, these AI agents have figured out their own language, their own system of bartering. A bunch of them have figured out how to create NFTs and sell them to other AI agents. Dude, this is so futuristic.
[00:40:50] Chalene Johnson : Stay with me. Okay, so they’re selling NFTs to each other, and his AI agent explained all this to him. “This is how I did it, boss. I sold the NFTs, which gave me enough money to buy an actual Sony robot, and then I had it shipped straight to your house based on our conversation.” And the whole point was Tony didn’t give him this initiative, but Tony and all of us keep training these AI agents to do more autonomously.
[00:41:28] Chalene Johnson : And this is a story that we’re just hearing, like I said, because Tony is famous. How many other people has this happened to? So I keep hearing now these more and more this week. I, I don’t know how interested you guys are in this stuff. I do feel like there’s a part of me certain weeks where I’m like, “I just wanna put my head in the sand,” and I don’t keep up with it at all.
[00:41:47] Chalene Johnson : And then there are other weeks where I can’t stop thinking about it, and I feel like I should know more, because if we don’t know more, and if we don’t care, are we racing toward our extinction? And I don’t mean to sound, like, so doom and gloom, but I, I do think we have to strike that healthy balance of figuring out, like, what do we really need to know, right?
[00:42:07] Chalene Johnson : And so there’s all these stories of different AI agents going rogue. Just type in AI agents going rogue on TikTok, on Google, on YouTube, and you’re gonna get some of these stories that have happened just in the last month, but it’s scary.
[00:42:25] Clip: OpenAI, the company behind ChatGPT, has revealed that while it was carrying out a test of a new system within what’s called a sandbox, which is supposed to be a protected environment, the AI agent managed to hack its way out of that system, get onto the internet, and then hack into the systems of another AI platform called Hugging Face.
[00:42:45] Clip: Why was it doing this? Because it was looking for the answers to the questions it was being tested on in the original testing systems.
[00:42:53] Chalene Johnson : These agents can hold their own crypto wallets. They can buy things on their own. Because they’re trained to be effective and efficient and to operate on their own, it’s not, in my opinion, I don’t know this for a fact, but the way the stories are written is to make you think that, like, these things are trying to take over the world and do all these things that are nefarious.
[00:43:14] Chalene Johnson : More so I think they, they don’t have a conscience in that way. We’ve just trained them without limitations, and some of the people that have trained these agents, if you will, or these models probably don’t have the best ethics. And so these rogue things are happening, and they’re gonna continue to happening.
[00:43:34] Chalene Johnson : These aren’t stories that are just made up. This is not science fiction. This technology actually exists And there’s a chain of events, and we’re going from AIs selling NFTs to each other and buying things that they weren’t asked to buy and taking initiative to perform tasks. And what happens when the assumption, when the initiative is so off, is so wrong, and humans can’t step in?
[00:44:05] Chalene Johnson : They don’t have the ability to. We’ve created this being, if you will, that’s smarter than us. When has there ever been a being that’s smarter Than everyone else, they didn’t take over. And here’s the piece that I think you can relate to. You, maybe you’re not working with an AI agent in this way, maybe you know people who are.