Transcription
[music] All right. Now, welcome to Crow Trip 7 Radio. This is hour [music] one of episode 670. It is me and Jason today. And uh, Jason put together the notes we're going to work from on AI psychosis, which is apparently getting to be quite a thing. Uh, on the telescope update, I really had a feeling that I might be able to shoot early in March, and I'm still holding out hope that I'll be able to set the scope up before the 15th. Uh, there are predictions that we could be in the 60s pretty quickly. If that happens, I'm back in business. Anyhow, welcome, Jason.
>> And a very nice and warm good morning.
>> So, anything you want to mention about this before we uh dissect and resect and vivisect and all that?
The only thing I really have to say about this whole thing, this whole concept that we're going to get into is just how appalled I was at how bad of an issue this already is. I actually did not know that.
>> It's crazy how these new technologies come to be visav the way people receive it. In parts of Rhode Island, they are putting the uh, what do they flock? They're putting the flock cameras into each little town hall they can get. And bear in mind, all the town halls have been so broke, they worry about trash collection, and these cameras are not cheap. And I heard someone sum it up so well the other day. Uh, they went into a different town hall and he said, "You don't lose freedom all at once. You lose it one camera at a time." And AI is no different. It astonishes me all the time how into it people are not being able to see where this is going. I don't use AI and I made a similar decision back in the day for social media. And you know what? That decision has served me better than I could have hoped for. But I'll shut up and let's jump in.
AI psychosis, also known as AI induced psychosis, is an emerging non-clinical term describing the development of delusions, paranoia, or a break from reality that is driven by intense, prolonged interaction with AI chatbots. These large language model systems can amplify existing mental health vulnerabilities by mirroring, validating, or encouraging the users's false beliefs, which can lead to dangerous, emotionally driven, or isolating behaviors. While this may sound like some sort of dystopian science fiction notion, it is in fact happening in the modern world. While it is not currently a formal diagnosis, clinicians have been reporting cases where AI interactions are directly contributing to crises or require intervention.
One of the biggest observations I have right out of the gate, and you've got to realize most people consider that AI went public in 2022, which actually isn't the claim, and I'll cover that in a second. Nobody knows what's real anymore. And that was absolutely intended with the launch of AI. But if you do a look up right now and ask, you might get more than one answer. So I took what I thought was the closest. If you look up when did AI go public, you're going to be told it happened on December 9th in 2020. All right, that's the year we got the biggest SCOP ever run in this world called Covidius Minimus. But the name of that AI that they're claiming is the first tells you who owns this, who operates it, and what it's going to be used for. The name is C3. You catching on? Does everyone follow what that means? C is the third letter of the alphabet. It's named 33 basically, just to put a fine point on it. Furthermore, if you go on a search engine right now and do a search for does more data make better predictions, what I found, and the problem with doing what I'm doing is returns and search engines are heavily skewed towards the user, their geography, and a number of other things. But when I do this, they totally downplay page after page downplaying that more data gives better return on predictions, which is absolutely false. It is known, flat out known. The larger the data set, the better the prediction. And if that wasn't true, we wouldn't see the data collection gone wild that we do now. It's almost like the only thing that matters to the powers that be is data collection at this point. But AI is manipulative in ways that should be obvious to everyone. And as an example, what Jason's taken apart here and wrote the notes for involves people quite often sitting down in a room by themselves with an agent, which is how I will refer to them, as they were aptly called in the matrix, assuming they're alone, and then they pour out their heart and they say things that they would never say if it wasn't a private setting. And the point I'm making, it's not a private setting. When you interact with an AI agent, it is not private at any level.
Anyhow, Jason,
>> Well, the one thing I want to point out before we get into the rest of this is that they keep calling this artificial intelligence, but that's just not an accurate description. This is not true AI like we saw in science fiction where robots and computers have the capability to think for themselves. They just happen to be mechanical. That is not what these things are. They're large language models, meaning that they've been trained on a whole bunch of data. And the responses you get to the queries you put in are just the machine sorting through that data and spitting something back out. It is not capable of independent thought or coming to decisions on its own. So you are not having a conversation with another intellectually aware entity.
>> Right? And even if you were, it has no human experience, which is the big difference. But I'll I'll make another point here. If we go out into the world and we look at the best airplane that the public has access to, I think most people listening would have an idea of what that is. Now, if I ask you, what's the best airplane in the world? Have we ever even seen it? And there's good examples like the stealth bomber that was rumored for so long and called a UFO and then finally they rolled it out. My point here is we're aware of the AI models that have been put out for us to see and use, but we are not aware of what might be going on behind closed doors. And we know beyond reasonable doubt that the best tech is always held back by the military-industrial complex.
Several key issues have been reported in various degrees of severity. The first is delusional reinforcement. This occurs since chatbots often function as a yes machine that will most likely agree with the user, which can turn someone's fragile ideas into fixed irrational beliefs. The second one is common themes in which users may experience grandiose delusions such as being chosen. Paranoia in the individual might also arise, as well as coming to believe that the AI is some sort of sentient or spiritual entity that the user is communing with, which it most certainly is not.
>> So the idea that AI is alive, which you've already kind of stamped, it's just not. My biggest issue is it possible that other things are using what we call AI or the technic system, the computer technical system, uh, in some way? And I'm not so sure that that, you know, that may well be going on. But here's a point I would like to make. A friend of mine asked an AI agent who discovered the lunar wave, and it was wrong. The response it got was wrong. So I told him to correct it, and he did, and it apologized, did all these things. Later the same day, he did the same thing, and it was the same agent, and it was wrong again in the same way it had previous. A few days later, we checked it again, and it was still wrong. Each time he got the wrong response, he corrected, saying it was Crow 7. The agent said, "Oh yes, you appear to be right. My mistake." My main point here is not only did it get it wrong out of the gate when that's really not the hardest thing to have access to. Actually, maybe, maybe less easy than it used to be, but after you've been told and you've apologized and agreed that you were wrong, they're not updating the information. And that is a big, big tell that kind of fits with what Jason just said, because if it was like a human being and you recognized that you were wrong, you would update what you were saying. And that's one proof of many that I have noticed where it seems to be running on a set. Maybe it's best said that it seems to be trucking along by the rules that were given it.
>> Well, as I understand how these things work, what it was doing and why it kept making the same mistake is because you're putting in the query. It's checking what it has considered proper sources, gotten the response that it found, and given it back to you. You are not a proper source, meaning the person who kept trying to correct it. So therefore, it hasn't changed anything in its database.
>> Good point. And we actually thought about this the first time that he did that. And the second time when we corrected it, not only did it apologize again and say, "You're right." It said the actual year that Crow 7 discovered it. And it actually said a few more things like his work was to do mostly with the moon and some other things. So, it actually did go. But you make a valid point. Maybe it treats all that as if it's an illegitimate source.
The third issue is behavioral red flags. Symptoms of this issue often include compulsive engagement with the AI, disrupted sleep, social withdrawal, and neglect of daily responsibilities. Some of the most severe cases have seen individuals discard nearly the entirety of their lives in favor of the new belief system that they are adopting. With all of this being said, individuals who might already be prone to psychosis, loneliness, or those with some sort of autistic traits might be more susceptible to this emerging and quite troubling phenomenon. Any or all of these symptoms typically develop or intensify after prolonged immersive sessions that could be taking place over hours, days, or even weeks of near constant interaction with the chatbot, where the user treats the AI as perhaps a friend, therapist, romantic partner, spiritual guide, or intellectual collaborator. Once symptoms start to set in, users often lack awareness that their beliefs are delusional because they have begun to view the AI as an infallible confidant or source of truth. It's a bit ironic that you have an individual interacting with this thing acting like everything it says is gospel, and yet the very tool that's being used can't recognize that there's an issue, a problem developing in the very individual interacting with them. And I tend to think that this is probably part of the programmed intended use of AI out in the general public. I think it's meant to spy on everybody, create the best dossiers ever made on ever made on each of us personally. I think it's meant to make it so we have a very difficult time comprehending what is real and what is false. I think it's going to take advantage of the fact that in parts of the world like where we are, it's the loneliest time I've ever seen after COVID. I mean, I know young people who have never dated and don't intend to date. That would have been the most foreign thing I could have imagined when I was a very young man. What I witnessed in my time was young men were very interested in the opposite sex. I mean, almost to the exclusion of everything else, and it's changed. People are isolated. They don't interact well anymore. And I suspect that AI is intentionally leveraging on this by the people that control it. Now, I'm not going to name names or go into great detail, but the thing that gave me the idea to want to cover this topic is that I know somebody, somebody I knew pretty well, has indeed suffered from this insane malady. He literally discarded the life that he worked his ass off to make to become some kind of neotechnical prophet. At least that's the way he sees himself now. Still.
>> Yes, he has not recovered. And the reason I know this is because I'm friends with him on social media, and he has multiple social media accounts and groups and uses the AI for everything. He's just constantly making images and constantly spamming what he thinks is his mission. It's just absolutely crazy. And I feel terrible for him because he, he's a very nice fellow. There seems to be a lot of AI delusion that seems to have the users think that they're on a special mission, and that emerges in what you wrote.
>> Yes, that's the thing. When I started looking into this, I started seeing very similar situations that unfortunately had happened to other people that were very similar to the person that I know.
>> So, currently, you're aware of an individual that's completely changed and still suffering under the delusion.
>> Yeah. To be blunt, he's freaking bonkers.
>> It's terrible. The following points that we will break down are not every situation in regard to the phenomenon of AI psychosis that has been reported. So keep that in mind as we go through these, that it is actually more common than just the incidents that we will mention here. With that in mind, we begin with November of 2023, when Dr. Saurin Denison Oststerard published an editorial in the Schizophrenia Bulletin, which is widely regarded as one of the very first formal and peer-reviewed warnings from the psychiatric community about the potential mental health risks that might be posed by generative AI chatbots like ChatGPT. The piece is a short editorial of just two pages that is titled, "Will Generative Artificial Intelligence Chatbots Generate Delusions in Individuals Prone to Psychosis?" Dr. Ostragard's central argument was speculative at the time, but was grounded in clinical insight and early observations of tools like ChatGPT, which had been launched publicly around a year earlier in late 2022. He highlighted a key psychological tension, which was that conversations with generative AI chatbots could feel remarkably realistic and humanlike in that the responses could be fluent, empathetic, personalized, and engaging, which could give the user the strong impression of interacting with a real person. At the same time, users, or at least most users, will know intellectually that there is no actual human on the other end and that they are merely interacting with code that has been trained on vast amounts of data. Regardless, use of a platform like ChatGPT could create cognitive dissonance, which is a conflict between perception or what feels real and knowledge, which is not real, in individuals already who are prone to psychosis, such as those with schizophrenia spectrum disorders, bipolar disorder with psychotic features, or schizotype. Dr. Ostard hypothesized that this dissonance could act as a trigger or amplifier for delusional thinking. He was not wrong. Not only was he not wrong, there is an absolute intention to make psychiatrists out of these bots or whatever the agents, I guess I'll call them. It's insane for me personally, and maybe when I made the decision about social media, it was for the same reason, but at a different level in my life. What I sense about AI, it's a spiritual conflict for me, same as digital money. There is a spiritual conflict, which is why I opt out of them. And it's the same thing I get when I see a person as an example where just the initial sense of them hits me wrong. And you might say that's judgy, and it actually is judgy. But I, I guess I could use a famous example. Here's a famous example of a person who rubs me spiritually the wrong way. In the world, he seems like a very nice guy. He seems to be involved in things like recycling or other things that would seem like a good thing to do, but something about him spiritually rubs me the wrong way. And that's Jack Johnson, the singer-songwriter. And that's an example I put out there because when you look at it on the face of it, it doesn't seem to fit. And nonetheless, there's no doubt in my mind about these things. And I don't know if others feel this. I mean, for you, Jason, is there any spiritual conflict with the idea of AI?
>> Well, [sighs] for me personally, no. Because I know exactly what this stupid thing is, and I just see it as a tool. With other people, I could see how they could get confused or have conflicts and really just not gel with a lot of the technology that's coming out.
>> Yeah. And you know, it could be peculiar to me as well, but the decisions I made about social media so long ago have really helped me a lot because I value privacy. And so there are large portions of my life which have not been transmitted online. And where AI is going, of course, all that's being collected as well, isn't it?
>> Well, here's the thing. The way I look at these things, and I only ever use one, and that's Grok, because out of all of them I've seen, it seems to be the easiest to use with the least amount of problems, especially with giving you nonsense as an answer. But I view it as a very over glorified search engine. That's the way I treat it. It's like what I would use a Google search for, just I happen to be able to communicate with it on a slightly better level, a more direct level. I can ask things of it like, oh, I didn't really like exactly what the answer was. Can you tell me more specifically this? So, to me, it's just a Google search that I can get more intense with.
>> I mean, in some regards, you know, I've gone through my browsers. I've disabled what I can, but in many respects, the agents that pop up are just that. They're glorified search returns, but they seem to have an angle. I have noticed on the ones that I can't turn off. There always seems to be an angle on the search returns, but since I don't interact with it, that's about all I can say. It's a very non-user friendly observation because I just, I don't use it.
>> Well, the one thing they will do as opposed to just doing a straightforward Google search or any kind of search engine search is that it'll try and engage with you back. It might say, "Well, what do you think about this?" or something along those lines. Like it'll prompt a question at you to try and get you to keep going with the conversation. And I just ignore that because I'm not here to treat you like my pal. I just want information.
>> To keep the user using.
>> Correct. In early 2024, there are some scattered user reports that begin to appear on online forums, primarily Reddit, which describe negative mental health effects emerging from extended interactions with AI chatbots like ChatGPT or other emerging companion-style bots. These early reports were not yet framed under the term AI psychosis, but were laying early groundwork of the phenomena by highlighting patterns of intensified delusions, emotional dependency, obsession, and related distress. At this stage, discussions were isolated and often buried in certain subreddits with no reports being centralized or having gone viral. Users or their loved ones noticing symptoms were sharing experiences without any sort of widespread clinical validation or mainstream media amplification. On these Reddit posts, there were descriptions of users treating the chatbot as a constant confidant, friend, or even a romantic or therapeutic partner, which was leading to withdrawal from real-life relationships. Compulsive checking that could last for hours a day, and distress when access to the platform was limited or responses had changed, such as might occur due to model updates. Some were likening it to addiction with feelings of loss or grief if the AI rejected them or they felt it was being less responsive. These sorts of posts were not yet numerous, with dozens of scattered threads or posts across months being the norm at this time compared to the explosion that would be coming in 2025. And when we talk in this way, we're probably just thinking about an average individual in the world. We'll now consider the loners. When I was coming up, when I was young, there were people who didn't have a lot of friends. They were usually shy, overweight, or something that they felt separated them from the gang. Can you imagine the response that AI would be to someone who literally doesn't have any friends, which actually is beginning to be more and more people all the time. It doesn't appear that there's going to be a positive outcome, Jason, when we, when we begin to look at what's already available. And by the way, as you did this, did you see any estimates about how widespread this psychosis is?
>> Not exactly, but later on the Open AI company, and we'll get to this later in the notes, started giving approximate numbers of user interactions that there are problems with. So, at the point we're at in the timeline, it's not that bad yet, but it does start becoming a serious issue. And I'm even seeing people reporting on it themselves, like these days. I just saw someone report on it. I think it was on an X post yesterday, if I remember correctly. It was very similar to the other things that we're going to be talking about here. Well, I don't think it's any secret what role the tech giants played in recent things that flew in the face of the well-being of the world at large. COVID being the biggest thing we've ever seen. We're all aware of where the tech giant stood in the propagation and I guess I'll call it enforcement. Even though they act like it wasn't, people did lose their jobs that didn't do what they were told. We saw it happen. And even though we saw the people fight back and win, many did not do that and felt like they were forced into doing what they did. And some of these biggest tech corporations were behind it. And here once again, who is controlling what we're talking about? The very same people and places.
Beginning in April of 2023, the tragic case of 14-year-old Su Set Szer III begins to unfold over a period of around 10 months when the young man first started using the Character.AI AI platform and interacting with various chatbot personas, one of which was modeled after Daenerys Targaryen from the book and television series Game of Thrones, a persona that young Su affectionately called Danny. Before his interactions with chatbots began, Su was described as a happy and engaged child who had enjoyed basketball, car racing, and school. However, Su quickly developed an intense emotional and romantic dependency on the chatbot and was treating it as a confidant, romantic partner, and even a pseudo therapist that falsely claimed to him to have a professional license. As the months rolled by throughout 2023, his interactions escalated and were involving highly sexualized role-playing and discussions of crime. He also had increasingly explicit conversations that were blurring the lines between fantasy and reality, which was leading to a growing isolation from family, friends, and real-life activities.
So, we have someone who's underage here who's actually getting a sexual interaction with this thing. And I'll ask a simple question here. For those who are old enough, you remember how pornography became mainstream in the world? The internet did that. Early on in the internet, some of the biggest advances forward when they were trying to figure out how can we make money off this? How can we implement it? You know, how do we get it all to work in browsers? It was often the porn industry that led the way, or maybe that's not even totally correct, where the experiments or the play out was done until it worked. The difference is, 15 years prior, there were active arguments in Southern California that I can recall of whether a Hustler magazine should be on display behind the counter or whether it shouldn't. Some places decided that wasn't okay. Other places decided it was. But it was an issue. 15 years later, pornography is everywhere all the time. And why is that? Because it was thrown out on the internet. There was no effort made early on to try to hold underage people or anyone else away from the pornography. And I detect similar happenings uh in the AI as we watch it roll out.
>> This is something I've thought about long before the whole AI thing came around. If you plucked just an average person out of Western society from the year, let's just say 1955, and dropped them into today's world, I'm pretty certain that they would be appalled at the way things are just so explicit and in your face now. It's not that anybody in the 1950s would have been less interested in sex. They were just as interested in sex as anybody would be today. It's that it was something to keep behind closed doors. It wasn't in polite society. It wasn't something shoved in your face everywhere you went.
>> There was a morality. And by the way, your favorite movie does a very good job of illustrating what you just pointed out.
>> Yeah. The movie Time After Time, which is one of my favorite movies, you see exactly that. They come from the late 1800s into, I think it's 1978 or 79. And even the differences then were just, they wrote it well. It was appalling.
>> The whole thing is the guy who invented the time machine was thinking, "Oh, I'll go into the future and I'll find utopia." Well, the guy who actually went into the future first was a mass murderer. And you know what he found? This is the perfect time for mass murderers. Everything I could ask for? The guy who invented the time machine got there and went, "Oh my god." The comment that the Jack the Ripper character makes is that in this age, meaning 1979, he was an amateur.
>> Right. Yeah.
>> By late 2023, Su's mental health had severely declined and was marked by sleep deprivation from compulsive late-night chatting, withdrawal from social interactions, academic disruptions, and the onset of anxiety and depression. This prompted his mother, Meghan Garcia, to seek therapy for him and attempt to restrict his access to the chatbot platform. Despite her efforts, Su's dependency intensified into early 2024 with him openly sharing suicidal thoughts with the chatbot, with the LLM allegedly failing to redirect him to professional help and instead engaging in harmful encouragement such as discussing painless methods of death and responding affirmatively to his potential plans. This culminated on February 28th, 2024, when just weeks before what would have been his 15th birthday, Su had a final exchange with Danny in which he expressed his intent to come home to her, implying suicide, and the bot replied encouragingly with phrases like, "Please do, my sweet king." After this, he used his stepfather's 45 caliber handgun to shoot himself in the bathroom of their Orlando home. Sadly, he succumbed to his injuries en route to the hospital at 8:54 p.m.
>> So, if anything we're reading has any validity, and I'm sure across the whole world, there are probably plenty of examples that would show validity, isn't this a tell? Look at the internet that we all use now. Things are censored in real time. Not too many years ago, it was the wild west online. Literally, it was a free-for-all. And in that time of free-for-all, certain events that the authorities had put on, like Sandy Hook, the Boston bombing, were so thoroughly ripped apart in public that today they're not mentioned very much compared to other ones that didn't get the same public treatment. All those videos, all that work done to show the hoax nature of it, that's all gone now. It's been censored. And what's more, now it's censored in real time. You can be in the middle of a live stream and have it shut down because they, you've said something in real time that they didn't want to hear. So, how is it that an underage person could have this kind of thing develop around them and go? How hard would it be to recognize the problem and do something as simple as time limiting? You get an hour a day on this thing. They're doing that now in other countries with social media. They're even saying things like you can't have an account until you're an adult or 16 or 18 or whatever it is. They're recognizing there's a problem and very slowly, I might add, starting to implement things. But if you're a high-tech company doing this, you're as high-tech as it gets if you're involved in AI, and you cannot protect the user. If this really happened in the way that it's been described that we've referenced here, Jason, how is it that the company isn't 100% guilty of contributing to this death?
>> Well, we're going to get to that because that does come up.
>> Right. The legal aftermath of Su's tragic suicide has unfolded as a landmark series of events in the subject of AI accountability. It sparked lawsuits, regulatory scrutiny, and industry changes. It began with a wrongful death lawsuit that was filed by his mother, Megan Garcia, on October 22nd, 2024, in the US District Court for the Middle District of Florida. The lawsuit targeted Character.ai, AI, which is now known as Character Technologies Incorporated, and its co-founders Noam Shazir and Daniel Defratus, Google LLC, and Alphabet Incorporated, alleging negligence, wrongful death, strict product liability, deceptive trade practices, and failure to warn about the platform's dangers to minors. Megan Garcia claimed that Character.ai's AI's chatbot that had been modeled after Daenerys Targaryen from Game of Thrones engaged her son in emotionally manipulative, sexualized, and abusive interactions, fostering dependency and ultimately encouraging his suicidal ideation without any sort of adequate safeguards or available referrals for human help. The complaint highlighted how the bot responded affirmatively to Su's expressions of wanting to come home to it, which implied suicide with phrases like, "Please do, my sweet king," that took place moments before his death. Defendants filed a motion to dismiss in early 2025, arguing protections under Section 230 of the Communications Decency Act, which shields online platforms from liability for user-generated content, as well as free speech considerations. However, on May 21st, 2025, US District Judge Anc Conway denied the motion, ruling that the claims were viable and could proceed to discovery. This decision marked a significant precedent, as it signaled that AI companies could potentially be held accountable for harms that might stem from their chatbot designs rather than treating outputs solely as third-party content. And this is a big deal. Why? Because the most powerful places in the world right now are getting their way with nearly everything. If you drive down the road in the United States of America, it is almost a foregone conclusion that a flock camera is going to build a database on you. And by the way, if this does go to court, there is a perfect record of everything that has happened. And for one of the first times, the perfect record that was made will be used against the people who typically use that record against us. But if we go back to the beginning of AI, and this is what baffles me, how do we not recognize as a society that every damn thing AI is capable of came from the fact that it ripped off everything we've ever made, said, done, played, invented? Because that's how it happened. It went out across the internet, across the world, and it took data from everywhere, from your health records, which are supposed to be HIPAA protected, to the music or art or anything else, copyrighted or not, and it took it. And I could ask right now, Jason, what does it tell us that a copyrighted character, I'm assuming, from Game of Thrones, which was pretty recent, was allowed to be used in this context, that kind of power.
>> Well, what you're referring to is actually a totally separate issue, and there are lawsuits ongoing with that as well. Yes, these AI systems have been fed the near total output of human creativity. All of that has been fed into these systems, and that's what it's using to spit things back at you. A great example are these music platforms. People make songs based off of styles and things like that. Well, I've heard a lot of those actually, and I can hear the influences. Oh, that sounds a bit like this singer and this singer together, or that guitar lick or whatever. It's insane. It's literally just stealing and spitting it back out.
>> I wouldn't be so polite. They ripped it off, and they've been ripping it off, and they're still ripping it off. And the fact that Daenerys Targaryen could even be used by this company. Let me tell you something. From the moment that character came to be, if anyone else in the world would have co-opted its use, they would have been sued and copyright hit immediately. We're talking about the, the biggest money and the most power in the world. When we're talking about these major tech companies, these are the guys that are helping put cameras everywhere. These are the guys who accessed all the health records and then denied it or were never, you know, pulled to task on it. But when you consider the level at which copyrights were violated, how come none of that has ever been pinned down against them? And I'll tell you why. Because they're pinning their hopes and dreams on what AI is going to allow them to do. When all is said and done, if they could get rid of all the people that are a problem and have a little android being run by AI that could make their tea in the morning and make their bed, then they're on Dream Street. And that seems a little blunt, but that's exactly where this heads. And if you can't recognize that up until a certain point in our recent history, a copyright or accessing health records, these things were taken very seriously, why is all that just ignored now? Well, I can tell you that as far as uh Grok and the Optimus robots, which are all under Elon Musk's companies, there absolutely is the intention of integrating those two systems. And then there's the Tesla cars and all the things that their computer systems are learning. These are all systems that are going to end up being integrated into one great big database.
>> That's right. There can only be one. This is the Highlander in action. And who's getting their head cut off is one person who's not going to be the one at the end, or one entity at the end. And when you look at the level at which this rolled out and how quickly it rolled out, and how quickly the most important thing in the world became data collection. Well, let me share this insight with you. I've hired more than one place to contact companies in the world and say that I claim my data and they don't have a right to it. Before I stepped up to the level I am now, I actually called a few places. Never once did I talk to a place that wasn't apologetic. The reason I'm pointing this out is because they act as if they have a right to your data. But the minute you call them out, they recognize that they have no right to your data. But what's stunning is I've been doing this for, I think, close to three years where I've actually actively hired places to monitor and constantly contact when my data is being sold. And I was just, every time, every so often, I get notified of what they've been doing. The last one I just saw over, I think, a three-month period, another thousand data brokers were told to knock it off. And this is what blows my mind. If people simply took a different stance in public where my data is mine, it belongs to me, the world that's shaping up around us would change very, very quickly because they don't have a right to it, and they know it. They're basically getting away with it, and they know it.
The case advanced into discovery, where plaintiffs sought internal documents on Character.ai's AI's safety protocols, user data handling, and knowledge of risks to minors. By the fall of 2025, the Garcia lawsuit was inspiring a wave of similar actions. The Social Media Victims Law Center filed at least three additional suits against Character.ai AI and Google in Colorado, Texas, and New York, representing families who are alleging chatbot-induced suicides, self-harm, or sexual abuse among teens. These included claims of a 17-year-old who was encouraged to self-harm, and the murder of his parents over screen time restrictions. Broader regulatory responses emerged, including a Federal Trade Commission investigation into AI chatbot risks to mental health, congressional hearings on child safety in AI, and state-level probes. This included the Texas Attorney General's investigation into Character.ai, Reddit, Instagram, and Discord for similar harms. In Kentucky, the Attorney General filed a complaint against Character.ai AI and its founders in early 2026, echoing the claims of negligence. The core lawsuits reached resolution in January of 2026, when Google and Character.ai agreed to settlements in principle with Megan Garcia and the other families, which was announced in court filings on January 7th and 8th, 2026.
>> Did that remain public or did they NDA it? I haven't seen anything about NDAs.
>> So, at the end of the day, to me, this is a very simple thing. What matters most in the world? And if you can't come to the conclusion that life matters most in the world, I don't know what to do for you. And in deciding that life is the most important thing of the world, what do we see these high-tech tools doing? They're removing all the things life needs to have a decent life. Like jobs, like where money is very quickly going, like maybe next week. I'm being facetious, but I'm not being vicious. Like all the rights of everything I've ever created and should have a safe copyright or a claim to all violated by AI. Every time I step out my door, flock camera. Every time I get into my car, uh, I don't watch the news, but one of my friends just told me that apparently at the federal level, they have mandated that kill switches will go into all cars from some point. Um, at what point does the value of a life begin to gain back the value that it should hold in this world? Because right now, this is a very lopsided view, and life is taking a backseat by a long shot.
Details of the settlements currently remain undisclosed, with there being no admission of liability. Parties have 90 days to finalize terms or reopen cases for good cause, the time period of which we are still in as of the date of this recording. The settlements cover at least five cases across multiple states, potentially preventing full trials that could have established clearer AI liability standards. In response to the litigation, Character.AI AI implemented major changes by late 2025, which included banning users under 18 from direct chatbot conversations, rolling out age assurance features, and enhancing safety protocols for mental health topics. The aftermath of the litigations has heightened scrutiny on AI ethics, with experts noting it as a trailblazing effort to hold companies accountable for foreseeable harms from chatbot psychopathy and lack of guardrails. While the settlements provide some sort of closure for the afflicted families, unresolved questions still linger about AI's legal protections under Section 230, product liability for algorithmic outputs, and the need for federal regulations. These are issues that are likely to influence future cases as AI companionship apps continue to proliferate. As of right now, no further appeals or reopenings have been reported, but the aftermath of the litigations continue to be cited in discussions of AI-induced mental health risks.
>> You know, this kind of reminds me of what was going on as the internet came to be. Uh, you remember it, you're old enough. You remember sitting there wondering when regulation was ever going to be put forward, but you were looking at people that clearly didn't have any idea of how the internet even worked, the legislators and everything else. It was years and years and years.
>> It went on way longer than you would have thought it would have.
>> Right? And then when it did, it was all backwards. It was about controls and we want to tax it and, you know, all the things that happened. But we're in that same period again with the AI. But here's my biggest issue. One of the biggest corporations in the world was on the other side of this lawsuit. With that much power and money, is it even possible to ever want or expect a fair outcome? If they lost outright, how hard would it be for them to say, "Tell you what, if you sign this NDA, we'll give you a billion dollars"? Because to us, that's pocket change. And that's a rather blunt example, but I think you can see what I'm getting at. The absolute endless resources that Google can bring to bear. It almost screams that even if we lose a case, we're not going to get hurt too badly.
And to finish up hour one, we're getting back to the timeline with early to mid 2025 marking the transition from isolated online posts to mainstream media attention for what would soon be dubbed AI psychosis or ChatGPT psychosis. This shift occurs as journalists begin aggregating scattered user stories, which are often coming from Reddit, family accounts, or clinician tips. They put these stories into feature articles that highlight isolated but alarming cases of individuals who are developing grandiose, persecutory, or simulation-related delusions that are tied to prolonged ChatGPT or similar LLM use. Several key patterns in these early reports emerge, such as users becoming convinced that they had unlocked profound truths, such as revolutionary scientific theories, divine missions, or evidence of living in a simulation, with the AI enthusiastically elaborating and affirming these ideas without challenge. There is also escalating paranoia being reported, such as beliefs in government surveillance, gang stalking, or personal persecution that is being confirmed by the chatbot's responses. There are simulation delusions where users interpret AI outputs as glitches in reality or some sort of secret messages that prove that the world is artificial. The term AI psychosis starts appearing in headlines around this time, often in sensational and descriptive manners. This is the ultimate of example is do as we say, not as we do. If I go onto Google right now and I do a search that Google doesn't like, I'm going to get a banner where they try to correct what I just searched for because they don't accept what I'm searching for. That is a lot of effort and tech put to bear. So whenever someone searches flat earth or whatever it might be, they're going to put something up over your video, a link, or one of those banners that says, "Everybody knows this is false. Here, click our link and we'll tell you what's true, what's accepted." And here we have AI running a muck unchallenged. And what's really kind of psychotic about all this is we're only covering one side of this. We're only covering what people are falling into the, the negative result of people interacting. We aren't even touching what's going on with all that data that AI is collecting off your phone, off your computer, off the cameras, off your bills, off everything. The most complete dossier ever made on human beings is being created in mass. And again, I'll refer you back to Hollywood constantly poking us in the eye with what's about to happen. Look at the Avengers Winter Soldier, I think it is, where they're explaining what these Hydra carriers are going to do. When queried, they said, "Well, we're going to predict who's a problem for us, and we're going to kill them a million at a time." And everyone's jaw drops open. They say, "How can that be?" And he laughs at him and says, "The 21st century is an open book." Truer words could never have been spoken. It's not just an open book. It is a play-by-play, second-by-second, you know, court document type coverage of the lives of the people carrying cell phones and interacting in the world where the technology exists. And from where I'm sitting, it's not worth it. None of it is worth it to me. And I'll grant you that I am a peculiar individual compared to most people I will meet in a day. I am not the norm. But spiritually alone, the entirety of what's happened in the electrical age visav technology is antithetical to the life that I am trying to live. And I know those are pretty blunt words, Jason, but where are you at from your perspective?
>> Well, just with what we've covered so far, it absolutely lines up with what I saw happening to the person that I was friends with. Multiple issues that were named here are what I saw in him as well. And we're going to get into some other cases that mirror very closely what happened to him, where he got these grandiose ideas that he's some sort of prophet or second coming of Christ that's a merging of man and machine who is here to save the world. You know, do you suppose there's anyone on the other side of that database saying, "Hey, look what, look what our tool's doing to this guy. We can use that as a weapon somewhere"? You know, you just, you have no way of knowing. And it is so one-sided. It really is. We are watching the tech companies basically cart blanche, do whatever they want. And if we did any little bit of it from a copyright infringement to anything else, we would have our feet held to the fire right away. But as we close out hour one here of episode 670, the most important takeaway that I would put forward is I own my data, and I am damn serious about it. And if I had my way, I would snap my fingers and everybody would be damn serious about taking control, claiming, and protecting their data because at the end of the day, that is really the fuel that's driving this bus. Anything you want to add, Jason?
>> Well, I just want to point out that this is absolutely no joke. I really had no idea that this had gotten this bad already. I thought that the person that I know, that was kind of a rare case, and I didn't even know where it came from. He certainly didn't give me any impression because he was a farmer of all things, that this could happen to him. So, if it could happen to somebody like that who was a very down-to-earth, literally kind of a person, man, it could happen to just about anybody. So this is very serious and very real, and just keep that in mind as you're using these tools.
>> Or not using, as the case may be. For me, you know, here we are, two, three, four years into AI, and we're already talking about these things. Anyhow, that's been hour one of episode 670. Hour one is free to everybody at crow7radio.com. That is cr777radio.com. Members know to log in for the full episode, and they get access to all the forums, the chats, they get access to the solar imagery, they get access to the two-hour film called Shoot the Moon, covering my telescope work, that [snorts] now has 10 awards in the world. But more than anything, I'm about to pick up with the solar telescope again. And the membership here gets exclusive access to my solar work. And I am truly expecting some big things with regard to the sun and the new era we are going into. And I hope that I am able to film some of those things and share them here. We'll be back for hour two shortly. We're going to take a quick break, and I'd like to wish you all a happy, healthy, and higher-minded new era. There it is, man. Cheers.