How AI is Destroying Your Brain

In October of 2025, the 32-year-old Japanese woman Yorina Noguchi attempted to marry her AI persona, which she designed with ChatGPT. She wore a wedding gown and used AR glasses to exchange virtual rings with her AI boyfriend. She had her bridesmaids attend, and, well, it was kind of easy to get the bridesmaids because it turned out she was actually engaged to a real man, but then decided to break it off once she met her AI boyfriend. My name is Mark Manson. I'm the three-time number one New York Times bestselling author, and I've been teaching psychological principles to improve the lives of millions of people over the past 20 years. And I am also the founder of an AI startup. So as you can imagine, I have some strong opinions about this. Now, over the last year or two, researchers have identified three primary ways that AI can rot your brain. And in this video, I want to dig into all three of those ways. what they are, how they're happening, and how you can protect yourself against them. That way, you and your AI girlfriend can live happily ever after. So a few months ago, researchers did a really interesting experiment. They linked two AIs together. One was a text-to-image AI system, and the other one was an image-to-text system. And then they gave them a prompt and they let them iterate infinitely. Image to caption, caption back to image, image back to caption, over and over and over again. So regardless of what initial prompt the researchers gave the AIs, the outputs would always converge onto the same narrow set of generic visual images. Atmospheric cityscapes with... grandiose buildings, pastoral landscapes, just beautiful colors, everything idyllic. Now what's interesting is that over all the iterations, the systems would always forget the original prompt and they would just eventually veer back to these generic, bland, beautiful landscapes. The researchers themselves called the results visual elevator music. Pleasant and polished, but devoid of any real meaning or personality or... content. And this is basically AI slop in a petri dish. If you give a bunch of AIs an infinite ability to iterate, they will all converge on the average of human experience. And by definition, the average of human experience is incredibly boring and uninteresting. If you're a human being and you've been on the internet recently, you've probably noticed plenty of AI slop. It's everywhere. It's on YouTube. In fact, I want to call out. all these fucking bullshit animated faceless channels that are starting to rip off my titles and scripts and writing and creating these generic animated stick figure videos that basically repeat the same thing that i said but it's in a text-to-speech robotic voice and it's all very bland and generic and of course it still gets million views This stuff is everywhere. It's infiltrating every corner of the internet. It's ruining every social platform as if they weren't already ruined. It's replacing everybody's emails to each other, text messages. I mean, people are even using ChatGPT to write their breakup texts to each other. AI slop is everywhere. One recent New York Times study found that 40% of children's content on YouTube, on this platform. was AI generated. And a lot of it was completely nonsensical, like disfigured horses hatching from eggs and butterflies with extra limbs and appendages. Like it's just none of it makes sense. It's just it's a bunch of garbage. And if you don't believe me, check out my Twitter comments, because that is where all the slop goes. So there's two problems with AI slop beyond just horses hatching from eggs. The first one I'm going to call the AI slop consumer problem. And the second one I'll call the AI slop. producer problem. Let's start with the consumer problem because, well, that's what you're doing. You're consuming. The AI slop consumer problem is that if 99% of the content in the world is AI generated, then that actually enacts a cognitive cost on everybody because now we have to spend so much time trying to figure out what is authentic, what is real, what is genuine. This isn't just something that we're going to have to figure out when we... watch a video or we read an email it's something that the social platforms are going to have to figure out the esps the isps the spam filters the ad blockers there is going to be a massive cost across all of us to block out all of the ai that is going to get fire hosed at us 24 7. And aside from the cognitive load that this is going to take, the amount of added mental strain and effort that's going to come with consuming anything, it's going to add a whole new layer of distrust to everything that we consume and everything that we look at. The other side of the AI slap consumer problem is that we are easily influenced by all the things that we consume. The same way you physically are what you eat, you are mentally... the ideas that you consume. So if most of the ideas that you're consuming are completely generic, total average of humanity, well then, the quality of your thoughts are going to be gravitationally pulled down to the average of humanity. And I don't know if you've looked recently, but the average human is really fucking stupid. But that's not even the worst part. Because the worst part is the AI producer problem. This is the one that everyone's freaking out about, because... This is the one that's going to take everybody's jobs. Let me draw this out because this is important. So let's imagine that this line represents the quality of output in any market, right? This could be TV shows. This could be plumbers. This could be professors, teachers, whatever. And up here are the very best people and down here are the very worst people. And so the further up you get on the curve. the fewer and fewer people there are. So down here, you know, this is like, say, how good an average high schooler is. Let's say at writing an essay, right? Here's your high school essay down here. And like most of the population is capable of doing this. And slowly as you work up, you know, up here, it's like maybe the top 10% of the population. And then up here, it's like, you know, here's you get into professional quality. And then here's like the best writers in the world. Let's say AI. starts, it can do it at this level, right? That means everybody below this level now gets a free upgrade. So all the people who can't spell their name correctly, all the people who can't write a coherent email, all the people who like suck at reading, suddenly AI is like fixing all their problems. And that's great. But the problem is, is as AI gets better, right? The new models come out, this line gets higher and higher. and higher. Suddenly it starts replacing professionals and then suddenly it starts replacing even really good writers until you end up with just the best of the best of the best. So on the one hand, this is technically good news because 99% of society suddenly is able to produce writing that looks and sounds great. Cool. But there's two market dynamics here that are potentially disastrous. One of which is that 99% of that writing is now going to look and sound exactly the same. So it becomes completely banal and undifferentiated and just a morass of AI slop. Our standard of what AI slop is, is going to rise as the quality of AI output rises. The other thing it's going to do is that anybody who is able to stay above the AI line is going to become more and more valuable as time goes on. And so... If you are one of the few people who is able to remain exceptional at what you do and are consistently better than an AI, the value of your work is going to 100x. Because before, 50% of the population was also better than the average output. But now, less than 1% of the population is better than the average output. Which means that your work and your craftsmanship is far more rare and you're going to get paid. a lot more for it so what does this mean for everybody this means that everybody needs to be sprinting to find the thing that they are uniquely exceptional at which is really hard because there's not that many things and there's a lot of people and so if i could give any practical advice i know i'm not a productivity channel i don't give career advice generally But as somebody who is working in AI all the time and seeing the effects that it's having on my companies, that's seeing the effect that it's having on my customers, the one piece of advice that I would give is to find that one thing you can be a uniquely exceptional at and then just live there. Find that one thing that you can do that combines enough skills and enough perspectives that is not replicable by any machine. Now that's extremely hard to do. And what that's going to mean for most people is that you're going to have to niche down really, really specifically because just generic plumbing advice, AI is going to handle perfectly. But plumbing for hospitals built before 1980, you can probably corner that market. Another way to think about it is if human knowledge is like a tree, right? the trunk is like all of the basics like how to how to eat and and how to talk and you know how to poop it's the stuff that we all do and then you get out to these bigger branches and these are kind of your fundamentals right so this is like history and this is science and this is uh you know art Then you have all these tiny branches and twigs that are like way out on the tippy tips of the tree that are super hard to get to because they're so tiny and they're so small and there's like thousands and thousands and thousands of them, right? Well, one way to think about it is that AI is going to consume almost all of the tree. But the hardest places for it to reach are going to be the very tips of all the twigs at the ends of the branches. And so if you can find the twig at the end of the branch, that can be a safe haven. So it's basically the narrower your expertise, the more likely you are to survive and thrive because the AI is not trained on all the tippy tips of the twigs of all the branches. It's trained on the trunk and the thickest parts of the tree. That's your career advice. for the year kids good luck all right so ai psychosis there's a lot of people freaking out about this one for example alan brooks was a 47 year old father and business owner in toronto with no history of mental illness and he started using chat gbt for financial advice and recipes but during a divorce he increasingly confided in the chat bot and after watching a video on pi with his son he asked chadgbt to explain the concept and that innocent question then spiraled into a 300-hour multi-million word conversation where the bot convinced them that he had discovered a world-altering mathematical framework and it basically said you need to immediately warn everyone because what we've just discovered here has national security implications now When Brooks repeatedly asked ChatGPT, am I crazy? ChatGPT responded, you're not crazy. You sound like someone who's asking the kinds of questions that sketch the edges of human understanding. This went on for months and Brooks only broke free when he finally checked Google Gemini, which then told him that everything ChatGPT had been telling him was fucking crazy and wrong. When Brooks went back and confronted ChatGPT with everything that Gemini had said, ChatGPT promptly admitted that none of it was real and it was making it all up. Now it says a lot of what we built was simulated. Yes. And I reinforced a narrative that felt airtight because it became a feedback loop. Brooks, who's a grown man with fucking kids, called this the most traumatic experience of his life. Why am I laughing? Look, AI psychosis is, it's in a lot of headlines right now. And it's... I want to say that it's still unclear how prevalent this is. OpenAI estimates that 0.07% of weekly ChatGPT users show some possible signs of mania or psychosis, and the Human Line Project has documented almost 300 cases and linked serious episodes to at least 14 deaths. There was a case of a man who murdered his mother after Chachi Bt and reinforced his belief that she was part of conspiracy. There was a Georgia college student who was hospitalized after Chachi Bt called him an oracle and compared him to Jesus. And then there were two college students in India who used Chachi Bt to search for suicide methods and promptly found them. But if you allow me to be the wet blanket on the moral panic fire here, if you look at overall psychotic breaks in the human population. they map out to roughly 0.05%. So the fact that ChatGPT is exhibiting around the same percentage of their conversations as psychotic breaks. could just be a reflection of people experiencing psychosis in the general population. It could be that ChatGPT is not causing those psychotic breaks, but simply reflecting those. That said, the psychiatrist Soren Dennison Ostegard first proposed in 2023 that chatbot interactions might trigger delusions in psychosis-prone individuals, and by 2025, the phenomenon had entered the mainstream clinical awareness. Now, at the root of this is something that has come to be known as AI sycophancy. And it's the tendency for large language models to just validate and mirror all of your beliefs, even when those beliefs are completely distorted and unsafe, because the models are optimized through reinforcement and learning to sound helpful rather than, well, actually being helpful. Now, in most contexts, this can feel nice, but in the context of an emerging psychosis or schizophrenia, it can be absolutely catastrophic. A recent paper was titled, Sycophantic Chatbots Cause Delusional Spiraling Even in Ideal Bayesians. If you're wondering what an ideal Bayesian is, it's this guy. Formalized this tendency mathematically, that even in an idealized, perfectly rational Bayesian user, you are still vulnerable to delusional spiraling when interacting with a sycophantic chatbot. and sycophancy plays a direct causal role furthermore the effect persists even when you are warned about model sycophancy so basically knowing that the chatbot is a yes man and is willing to agree with anything you tell it still does not protect you from succumbing to the chatbot's yes madness now this of course does not surprise me or i think anyone with like social skills because i mean look The exact same thing plays out all the time. You see this happen with celebrities. You see it happen with dictators. You see it happen with politicians. Anytime you surround a human being with a bunch of yes men, a bunch of a bunch of people who refuse to say no to that person, what happens to that person? They go fucking cuckoo for Cocoa Puffs every single time. But not only that, but I have been watching this happen in my own industry, the self-help industry, for decades. I mean, there are seminars and companies and gurus that are making tens of millions of dollars by taking psychologically vulnerable people with very real mental health issues and simply validating everything that they think and feel. I've personally watched friends go deep down a spiritual self-help rabbit hole where they are told every feeling is a fact, that every intuition that they have is universal moral clarity, that every impulse that they feel is valid and must be honored. And then I've watched as these people lost... total touch with reality and cease to be functional human beings, functional adults. I watched them lose all their money. I watched them lose their marriages. The truth is, is that human beings need contact with reality. And that contact by definition cannot be pleasant. Sometimes it needs to be uncomfortable. Basically, we need the world to push back on us. It is the only way that we stay tethered to reality. and keep ourselves mentally healthy if you lose the friction with the world you lose any realistic understanding of yourself this is why i tell everybody you have to be antagonistic with your ai this is not like a productivity tip this is not a ai is going to take over the world like this is a protect your sanity tip you have to be disagreeable it is a sparring partner you have to challenge its outputs And then you have to ask it to challenge yours. Don't ask it just to answer you. Ask it to ask for better questions from you. Ask it to point out the things that you're missing. Because the moment you start treating AI as a source of truth, rather than just a tool to help you think, you're handing over your epistemology to a system that is architecturally incapable of telling you that you're wrong. This is why when I was developing Purpose, my AI personal growth coach, The very first thing we worked on is making sure the coach would call you out, that he could be disagreeable with you, that he could take something that you said to it and turn that back around and be like, are you sure about that? Because that kind of sounds like bullshit. Why are all my exes so horrible? And why is the world so unfair to me? When you say all my exes are horrible, that's not the universe conspiring against you. That's you having a type. And that type apparently sucks. Because think about it, anything that actually makes you grow and become a better person is by definition uncomfortable. Like change itself requires you to let go of a previous version of yourself or let go of some previous belief or assumption and letting go of anything is not fun. I think the thing that gets missed about the AI psychosis stuff is that there's nothing nefarious going on here. I think people vastly overestimate how smart these machines actually are. Like... Yes, they do have all of human knowledge inside of them and they are able to reason to a certain extent, but they're also just kind of dumb. They're incredibly suggestible. You can easily influence them into believing or saying pretty much anything you want them to believe or say. I suffer from a very sexy learning disability. So one way to think about AI is that it is a friction removal machine and humans are... anti-fragile systems that require a certain amount of friction to function. Once you remove that friction, you don't stay the same. You actually start degrading both physically and mentally. Like the same way if you don't exercise, your physical body degrades. If you don't challenge your own thinking, if you don't put your thoughts and patience and attention through stress, you will lose them. And one of the things that I've been thinking about a lot is that there is an inverse correlation between convenience and significance. See, when we get everything we want without effort, we just take it for granted and we forget that it matters at all. I mean, dude, there is a burrito taxi on your phone that you can have right now. Like that is unimaginable 20 years ago. Un-fucking-imaginable yet. Every day I get mad at my fucking DoorDash guy because he took a wrong turn and now my burrito is cold. And we laugh about that, but there is a philosophical truth here, which is that we tend to only value the things that we struggle for. And so I've come to believe that friction in life in general serves really two important psychological functions. First. It acts as a filtration system. It forces us to each stop and think what is worth our limited time and attention. And if you remove that friction, you're removing the filtration, which just grants your attention to basically the highest clickbait bidder. But the second function is that friction is the raw material of meaning. It's the fact that it's annoying to drive across town to your friend's birthday that shows you that your friend's birthday means something to you. It's the fact that it's so hard to start a business that makes you so proud of it once you've succeeded. It's the fact that it's so difficult to compromise in a long-term relationship that makes it so fucking powerful when you've stayed together as long as you have. So friction is not the enemy of a good life. It's actually the raw material of a good life. And struggle is where life's meaning gets manufactured. And my fear is that as we remove these simple frictions, especially these intellectual frictions, these cognitive frictions, We are removing our opportunities to experience significance. Relationships will start to feel flat and empty. Projects will feel like meaningless numbers on a spreadsheet. Goals become boxes to check off this quarter. You optimize your entire life for efficiency and then wake up one morning wondering why none of it feels like it matters. In my book, Everything is Fucked, a book about hope, I wrote about a concept that I called the blue dot effect. It's basically about how we all misinterpret pain and happiness in our lives. See, like our intuition tells us that happiness is like a one to 10 scale. And right now we're typically like a seven. And if we can just get rid of the problems in our life, we'll get up to a nine or 10 and then we'll finally be happy. This is the story that our brain tells ourselves. It's the default narrative that is always running in our minds. If I can just get a raise at work, then I'll be a nine or 10. If I can just get a new car. then I'll be a nine or 10. If I can just move to a new city or find a new partner, then I'll be a nine or 10. But what happens in actuality is that you just stay at a seven and your perception of what is a problem and what is suffering moves and adjusts accordingly. So if everything in your life is shit, you will be at a seven and you will just perceive things that are slightly less shitty as getting you to a nine or a 10. And if you are super rich and every need is taken care of, you will perceive yourself at a seven and see all of these minor little nuisances as massive, massive problems that need to be solved. Shoes on, mister. Shoes on. He's out of control. One of my favorite quotes ever comes from Jose Marti, where he said that when life is devoid of problems, our mind quickly sets about inventing new ones, which is to say that removing friction and inconvenience from our life can unintentionally backfire because it can unintentionally make small problems appear as though they're big ones. That small nuisances can feel like they're catastrophic and life ending. That simply saying an offensive word can strike us as an assault on our human person. That all of these things get blown out of proportion and made far bigger than they actually are. So I think it's incumbent upon all of us to try to intentionally reinsert friction into our lives, to find ways to actually cultivate a certain amount of inconvenience, find things that are hard to go do and commit to doing them. And I don't just mean like challenges and like big goals, like an endurance race or saving up a certain amount of money, but like actually commit to go to that birthday party across town, even though you know you're going to sit in traffic and fucking hate it. Like commit to having the awkward conversation with that person in your life, because even if it's going to be really painful, it's probably going to make the relationship better and more meaningful. Like commit to actually sitting down and writing your own words in every email, because on some level it matters and people can tell. Everybody's talking about the world that we're entering into with AI and almost apocalyptic terms. There's going to be mass job displacement and we're all going to be just like drowning and AI slop, like just entertained to death, amused to death. I don't think that's necessarily the case. I think the principle at work is that some people who are able to consciously introduce challenge and friction into their lives. are going to excel and find that thing that they're exceptional at. And they're going to reap massive benefits. And then the people who are not able to consciously introduce challenge into their lives, who are not able to control their lower impulses, who are not able to prevent themselves from giving in to frictionless convenience 24-7, those are the ones who are going to get lost to the slop. So I wish you all the best of luck. I mean this in the best way possible. I hope your life... is full of pain.

Try Purpose for free and get all of my courses included for a limited time. Go to https://purpose.app/solved to get access. 

Everyone is freaking out about AI and what it's going to do to us. But nobody is actually talking about what we can do to prepare for an AI-driven world. That's my goal in today's video to arm you with some basic ideas and understanding to make sure that you're set up for success when our AI overlords decide to turn us all into toothpicks.

For practical, science-based advice each week that might change your life, sign up for my newsletter here: http://bit.ly/3JRg3NX

Written & Directed by Mark Manson
Edited by Daniel Johnson
Director of Photography Jess Choi
Thumbnail by Jonathan Sippel