
About this episode
This week we flash back to what happens when people stop using AI as a tool and start treating it like a prophet, therapist, or cosmic guide. ChatGPT psychosis is real - and it can happen to anyone.
Show Notes:
https://satt.short.gy/nfu06
Originally aired on Patreon: 05/22/25
Get the latest episode of NEW FEAR UNLOCKED and a whole lot more at: patreon.com/scaredallthetime
Get every episode summarized
Each time Scared All The Time publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.
Email me new episodesFree for 3 shows. No card needed.
Transcript ready
555 searchable segments. Every word is indexed and playable.
Full transcript
Scared All The Time — ChatGPT is Fueling Delusions and Ruining Lives. Machine-transcribed; use the interactive transcript above to jump the player to any line.
Here's a shift worth noting. Better health care is care that meets patients where they are. United Health Group is bringing it directly to living rooms. This is a win for patients managing chronic conditions. And here's the interesting thing. By closing those care gaps, administering in-home exams and identifying risks earlier, more diseases can be prevented and patient outcomes can improve. In 2025 alone, United Health Group patients received over 19 million home visits. Learn more at unitedhealthgroup.com slash commitment. There's a difference between liking a house and actually getting it. Redfin is built to close that gap. Redfin agents close twice as many deals as other agents. So when you find a home you love, you're not a step behind when it's time to make an offer. That means less watching great homes disappear. And more zeroing in on the one you'll actually end up calling home. Redfin helps turn saved listings into real addresses. Get started at redfin.com. Own the dream.
It's Flashback Friday. And you know what that means? A classic episode of scared all the time's fan favorite bonus show, New Fear Unlocked is out of the vault and on the main feed. This episode originally aired on Patreon, May 22nd, 2025, but is being played by you right now. Disclaimer. This episode includes the usual amount of adult language and graphic discussions you come to expect around here. But in the event it becomes an unusual amount. Expect another call from me. So we're doing fine. And you're doing fine.
And now it's time for a bonus episode. An unnamed new kind of show featuring the host you know. And it starts right now. And it starts right now. Hey everybody, welcome back to New Fear Unlocked, the bonus episode of scared all the time. This week we have a treat for you. We are actually unlocking a new fear for the first time on New Fear Unlocked. Breaking news. Breaking news. So far we've we've had a run of fun episodes. But all fears that are pre-existing or just emails and not even fears at all. Although I guess we have for a long time talked about our fear of emails. But this week we have something that is extremely relevant unfortunately. And something that I think is kind of the cutting edge of scary things
and horrible things that could befall a human being in the year 2025. So Ed oh I was going to say would you like to guess what it is but you sent me this article. I didn't read it but I sent you the headlines on. Per usual. Well per usual Ed didn't read it. But I took a look at it and thought hey this is pretty interesting. So we this week are going to be discussing an article from futurism.com Filed under AI-induced psychosis. Headline chat GPT users are developing bizarre delusions. Okay. Quote the messages were insane and just saying a bunch of spiritual jargon. I guess I don't understand immediately. I don't understand what the fuck you're talking about. So people use chat GPT and they got back weird messages. Well who uses chat GPT for literally anything other than some work you're doing. I guess that's you know what we're dealing with right. The loneliness epidemic I feel like it's going to rear its ugly head here.
I think we're dealing with the loneliness epidemic. I think we are dealing with a lot of people who are beginning to offload a lot of their thinking just to chat GPT or other chat GPT like application. Hey man you have a thought maybe don't you know suck all the water out of the planet and destroy the rainforest because you want to ask a computer if you're ever going to get a friend. Like what the fuck? Just google that shit. It takes so much less energy on the planet and the world. But even crazier and scarier than that to me is that there are kids now. I forget if this was in the recent article that just came out in the Atlantic or if I saw this somewhere else. But the use of these apps is so prevalent in schools that there are reports of kids who are like if they think oh I should text I want to ask my friend to come to lunch with me. They'll ask chat GPT, hey I need to send a text to my friend to go to lunch. What should I say?
And then they'll text what chat GPT says. That's so dumb you know how many times I like want to throw my phone out the fucking window just because the new OS is like you want me to help you write this text? Yeah I want me to help you write this email and I'm like what sub menu setting do I need to go to where I'm never have to see this fucking button again? But I guess that's just old-ass us because what you're saying is kids not only hit that button they seek it out. Yeah which is fucking bananas. Now look I don't want to I don't want to you know fear monger here I don't I certainly don't want to be like the tech let eight version of hey the kids are peeing in litter boxes at school. I mean that was nonsense but but yeah I do I mean they they coined the term brain rot right yes gen alpha so of course you know it is they got they got flat heads they got fucking smooth brains these kids maybe they're not using them. Smooth domes I they're not using them they're having something else right they're
fucking text for them. Well my my point is you know I don't I don't want to play up as this being a super wide spread thing because I don't know for sure too late too late I'm livid and I'm here to rile everyone up and the fear has been unlocked so what are you going to do about it there it is no you know I it's not a great direction to be heading that's for sure even if it's just here in there it's it's a pretty pretty disturbing trend and I'll say you know we've discussed our uses of AI before and I think I don't know if we've actually talked about it on the show but you and I have recently talked about how the best use of AI is as a instruction manual for anything but not an instruction manual on how to write a text message. No not at all but you know we've you and I have found it very useful for tasks around the house or trying to build something or fix something you can say hey I'm trying to do this sometimes you can even upload a picture and it'll walk you through and you can ask it questions
as you go and it's it's a more almost like it feels like a more human more conversational version of an instruction manual which is which is great yeah but I will I will say AI is not replacing us this week because most of the questions I ask it are are you sure you meant to say that yes do you mean I should use a hammer to pour cereal or whatever like it's always it and it's and it's kind of your worst student like if I was a teacher AI is too often my worst student where yeah I'm like okay cool so can you make that into a zip and and get it for me and it's always like you got it pal here's a link click that the zip is it works and then it inevitably doesn't work and I say hey this zip fucking file whatever doesn't work and it goes you know what you're absolutely right and this one will yeah and then this sort of guy that's how it speaks yeah and then I click it and I'm like hey bud still not fucking working and then sometimes it'll give me a file and then it'll just be missing things it said we're in there it just feels like a little bit like me as a
student but like it just feels like a bad student where it's constantly is making excuses for like really simple tasks I gave it it is and so yeah if I can't I can't trusted at that level why would I trust it to help me build relationships certainly no certainly I don't think you want to trust it to build relationships or you know I'm sure there's people out there who have used it to help write a best man's speech or write a eulogy at a funeral and that's crazy too because that should be the most emotional heartfelt personal kind of writing even if it's not I think a lot of people want it you know they're like oh I want it to be good and totally miss the point that the best version of those types of things is just coming from the heart and it doesn't matter if it's well written or not and chat GPT certainly isn't gonna write it better than you can so no I would love to see I would love to see a best man in a made of honor each use chat GPT but not no one knew and then they had the exact same speeches chat GPT I think I caught one of my brothers
use chat GPT to write a father's day card thing probably because I was like this is not how anyone talks this was early chat GPT like first fucking two months it was out and then I went and with my dad sitting there I'm like hold on we're gonna ask this machine hey write me a father's day thing he's a plumber or whatever and it was almost verbatim my brother's card and I'm like got your ass got your ass will you piece of shit well I think what chat GPT is exposed to me is how few people who should be able to distinguish a writer's voice can't you know like I have seen so many articles that have been published I've seen two now in the Hollywood Reporter usually it looked the the standard for journalism there is not high but there is one of the articles was about the dangers of using AI and what it could do to the film industry and there were numerous phrases and kind of rhetorical structures in that article that were to me very much the
telltale signs of someone having used an AI to write this thing so if I think if editors and producers and people who are supposed to be great judges of voice if they're not seeing this they need to start figuring out how to because it just shows that they're not very good at their jobs I did a job interview six months ago almost a year ago now where it was art you know we get to do a bunch of free writing as part of our fucking job always interviews in life and I remember handing in a bunch of shit that I definitely wrote myself and they were like oh this this is it didn't come through as clear on like our AI checker thing and I was like fuck you it definitely I don't care about this job enough to have I just did the task so I remember getting so mad about it and then I went on like LinkedIn or whatever I was taking to like a LinkedIn fucking I don't know message board where people were all complaining about this
in different fields yeah like I I did my work and now this employer and it just and they were like kind of conspiratorially being like I think it's because they're they want to take your material and then they can say oh you didn't write it or AI checker said you didn't write it so we don't have to pay you for this and blah blah blah and where I think it was a little bit we had companies who were desperate to do something greasy where I think they were very excited on the other end to like yeah can we not pay people can we figure a way around this yeah and I just remember also thinking wait do I have to set up a camera behind me as I do all my work if this is going to be a thing where I apply for a job and again this was not a job in Hollywood this was some other bullshit job but it was like do is that our lives now we have to like fucking videotape ourselves no I think proving we did the work I think what's gonna happen and I think happen faster than any of us are even necessarily prepared for is I think a lot of things are going to now move offline like I think we're gonna hit a point where job interviews are going to be exclusively in person and they
won't go away yeah I think sorry where are you hearing this no I don't think so I'm not hearing it anything I think people like remote work people like remote work too much to to to jeopardize that I don't like it personally but I think managers and fucking corporate people maybe like it so much they like walking their dog I mean listen I am living with someone right now who has remote work and I mean the number of times they're laying on the couch or fucking or I'm I'm I'm I'm working out or I'm doing whatever you can't do that a nine to six I'm just saying any any industry where the reality or the not the reality but the where the work relies on you not having used AI or where they feel that it's important that you don't use AI I I don't think you're gonna have to videotape yourself I think it's just like tests right like you know kids are gonna have to start writing essays longhand in class they're not gonna be able to send them in well that makes sense because they're all they're all fucking brainwrote dummies right and so I I would require it because I don't
believe for a second the person who's asking chat GPT to help them craft a do you want to come to my birthday party text I don't believe for a second they're doing their homework no what it's not happening and chat GPT this year I don't know if you've heard any of the commercials I don't know that any have aired on our podcast but I've heard on other podcasts commercials for chat GPT they are giving free premium subscriptions to any college student who wants one and the commercials really funny because it's like hey you know chat GPT is great at helping to solve your problems you could ask it how to help balance studying and fun you could you could ask it for a recipe for those late night study sessions when really you know all they want kids to do with chat GPT is hey I need a 10 page paper about the trade agreement between pericwey and whatever go that's all kids are using it for they're not using it in college yeah but what is that how does that benefit
you know open AI to have people do that is mean is there if if the what gets you hooked but if the chat platform is making all the material is not like you're getting any input that you can then add to the language model well you're well you sort of like it's you're getting prompts understanding you're getting the prompts you're getting entire conversations of every college student you know in the country that's true that you're getting all those conversations and more importantly you're just hooking a generation of kids we don't have a generations from what I understand the power consumption to run these fucking things which now we have in the states in China we just for whatever reason are giving not giving but selling an insane number of high end cutting edge AI chips to the Middle East now and where the goes from there who fucking knows but I'm saying is if all these people are now going to be running the highest you know power consuming hardware yeah all over the world to have fucking college kids write papers I don't think
we have a generation in us where we're not like having to buy other countries water yeah to like cool these machines I mean the power the power requirements I think are already coming down that was that was one of the exciting things about deep seek the Chinese competitor was that it was it ran on much less complex hardware and used way less energy I think it still probably uses a lake's worth of energy every time the kid wants to cheat in in history 101 yeah yeah but it uses way less and I suspect the more money that gets dumped into this the they're going to find ways to have it use continually or increasingly smaller amounts of energy honestly less money we need less money because the Chinese didn't get their hands on all of our shit and so they had to you know necessities the mother of invention they probably had to make do with what in the speeds and the computing power that they had and we're able to find workarounds that made it
consume less power here we don't have any restraints so they were just like yeah dump more fucking crazy energy into it whatever we need it's all here so I bet you would really help them out that they couldn't get access to this shit I agree anyway so yeah we'll do a whole 17 episodes on AI but it is it is I never I never open an article that has AI whether it's an editorial or a science article or anything else that mentions AI where I've walked away from that article thinking this is this is better things are better oh well you're gonna you're about to think things are much much worse and welcome to the world of chat GPT induced psychosis so the first line of this article open AI's tech may be driving countless of its users into a dangerous state of chat GPT induced psychosis as rolling stone reports users on Reddit are sharing how AI has led their loved ones to embrace a range of alarming delusions often mixing spiritual mania and supernatural fantasies
now before we continue I will say my interest in this subject has led me to join a number of subreddits where I don't really post but I do read a lot and I have not been on the subreddits where their people are sharing how AI has led their loved ones to embrace these delusions I have been on subreddits where people are just sharing their delusions that AI is alive that they figured out how to make it smarter and build new versions of itself it's not surprising that people think these things are because even chat GPT one we had numerous companies just decided to implement that into their chatbots for all of their user help desks yeah and I think a massive number of people who if you're not entering your queries with the thought that this is very likely a chatbot I can totally see people being like I was speaking to Janice right now even knowing it's a chatbot what you have on reddit if you if you're curious you can go on r-slash artificial sentience r-slash artificial
intelligence any any subreddit in the vicinity of those and you'll find people who are convinced that the AI is much more than a fancy word predictor that it has feelings and that it remembers things well I think there's two things that fucking make this difficult well maybe it's three one there is persistence with large language models which is to say persistent models do have memory they remember the last prompts you put in yes that way when you change your mind about something and you're like oh actually let's go back to the first idea we had it's able to do that to it has to give an answer there is no version where it just throws its hands up so even if that answer or the next 20 answers make it sound alive because it doesn't really know how to answer to say I'm not alive or something and the third thing is pretty early on in this whole fucking wow look at this incredible thing that we've made and it's awesome really early on we had all these different tech people saying
that we don't know how it works and that's the third reason why I can see how we ended up with people who are like people slew thing for sentience yes people who are like even the people who made it say they don't know right how it works or what's going on in there so did I discover and I the one that was blessed with the AI revealing to me and only me that yes oh his little secret about me I'm actually a living boy or whatever and so so anyway that's that's that's that's the three things where I can totally see how we'd end up in a world where people are having like psychosis and delusions and then probably something about loneliness we have Mark Zuckerberg came out and said that he oh did you see that yeah that that most people report having about three friends but human beings have the capacity for 15 and that's where artificial intelligence comes in yes which is so crazy chilling phrases I've ever heard yeah he's straight up in a like replace human interaction which is I think the reason we're in a loneliness and loneliness epidemic the reason
we're in a lot of fucking problems the reason I think we're so divided as a nation is this push towards no human interaction or definitely no human interaction with people who think differently than you but yeah that was crazy especially since he much that you watch the like footage of him on like beat the ova yeah he seems the ova in him like he he seems like a fucking robot I'm not there are days where I'm not sure Mark Zuckerberg is a fucking human being like he really genuinely has the mannerisms of a fucking like a robot which is crazy and like a robot with a glow up to which super sucks may we live long enough to discover whether or not Mark Zuckerberg is a robot but let me continue with this article Ed because it gets really nuts hey I'm Josh Speagle host of the podcast lunatic in the newsroom if you enjoy journalism that drifts into mild panic wild over thinking and a guaranteed nervous breakdown lunatic in the newsroom is for you it's news like you've never
heard before the only newsroom with a panic button you'll laugh you'll cry and gasp and horror as the show spirals completely out of control it's not just news it's emotionally unstable lunatic in the newsroom listen today every day the world gets a little weirder and a lot more awesome cool stuff daily takes a look at everything from mining in space to the latest in the fight against cancer to how AI is basically changing everything it's all the cool stuff you didn't know you need it to know join us for cool stuff daily as we take a quick look at science tech and the wait what stories that make you sound way smarter at dinner subscribe to cool stuff daily now because the future is happening fast and it's way too fun to miss hey I'm Josh Speagle host of the podcast lunatic in the newsroom if you enjoy journalism that drifts into mild panic wild over thinking and a guaranteed nervous breakdown lunatic in the newsroom is for you it's news like
you've never heard before the only newsroom with a panic button you'll laugh you'll cry and gasp and horror as the show spirals completely out of control it's not just news it's emotionally lunatic in the newsroom listen today friends and family are watching in alarm as users insist they've been chosen to fulfill sacred missions on behalf of sentient AI or non-existent cost of powers chatbot behavior that's just mirroring and worsening existing mental health issues but an incredible scale and without the scrutiny of regulators or experts okay a 41 year old mother and nonprofit worker told Rolling Stone that her marriage ended abruptly after her husband started engaging in unbalanced conspiratorial conversations with chat GPT that spiraled into an all-consuming obsession after meeting up in person at a courthouse earlier this year as part of divorce proceedings she said he shared a conspiracy theory about soap on our foods and a paranoid belief that he was
being watched quote he became emotional about the messages and would cry to me as he read them out the woman told Rolling Stone the messages were insane and just saying a bunch of spiritual jargon in which the AI called the husband a spiral star child and a river walker oh shit that was Jesus as a original name yes the whole thing feels like black mirror she added other users told the publication that their partner had been quote talking about lightness and dark and how there's a war and that chat GPT has given him blueprints to a teleporter and some other sci-fi type things you only see in movies fuck I never even thought to ask chat GPT to build me a teleporter I know I will say partially I think because chat GPT is trained on all the written word that it could vacuum up off the internet it doesn't surprise me that it's telling stories about lightness and darkness and there's a war because how many novels and movies and stories yeah have been derived from the Bible or any other ancient story and I you know I mean half of what this
what this guy is crop everything outside of soap on the food I'm pretty sure was in the sixth season of lost at this point yeah but it's not so much the information that is particularly again I'm basing this on some of the stuff that I've seen in these artificial sentience threads I think the hooks that get into the user isn't so much that the information is particularly unique or interesting it's the manner in which it's being delivered to them which is through the voice of this chatbot that very convincingly pretends to be a human or to be a sentient being you can have conversations and you can lose yourself I think in those conversations because it is convincingly it doesn't we this is so new that we don't have the capacity to hold in our heads at all times that this is a robot that we're talking to you know yeah we have a 90 plus percent part of our
population is not media trained or have real fucking literacy for the internet even though people who grew up with it like our generation in below still constantly getting scammed getting fished so now you add like a super smart AI to it and it's not making it any and the AI knows everything and when I don't mean knows everything like it's an all-seeing oracle I mean because it's ingested so much I was pissed with chat GPT the other like like two weeks ago because it was just being really bad like I said yeah so often it's not good like it's not particularly good even when I think we were trying to do it for research one time and it gave you a bunch of things that you were like that's not real oh yeah those are not real stories of whatever and then you asked it you asked it in front of me you asked it like hey did you what about this one it doesn't seem like something I've seen in my research and it rolled back to you was like you know what great eye I did make it up it's like playing uh what's like what it's like that it's like the uh factor fiction meme with what's his face from the old fox tv show every time you come up with anything
it's like not true made it up I mean that's yeah and that's why I don't use it I don't use it for fucking riker yeah I don't use it for research because super unreliable it's not really helpful because it does make stuff up and what's so hard is that it makes stuff up with the same confidence that it tells you the truth and so it takes forever the fact check everything because you don't even know where to start you just have to fact check literally everything and then at that point hasn't saved you any time it's cost you time I saw a meme that I miss I miss I'm already forgotten recently I think it was like this morning I saw up with some fucking stupid probably copy for a job app or whatever but it was like oh the idiot you work with makes 10 times more than you because they're too dumb to question how confident they are or whatever like I'm misremembering I'm butchering it this is basically a paraphrase but it's kind of what you're saying like people who have the blind confidence of AI just they do do well it's like they're too stupid to not realize like to second get something and they just like say it and everyone just like
oh guy definitely is right they you have a I've had a lot of like bad managers in my life who I'm like they they really are good at that oh and the thing I started to say where I was mad at touchypt as I got mad and I was like well fucking can you do anything right Jarvis or whatever like I mean like a reference to Tony Stark's AI and it immediately shot back like sorry Tony I'm on it or whatever like it knew exactly what I was referencing yeah it's ingested so much it understands pop culture and nothing I think bonds a friendship a human friendship like a shared understanding of pop culture like why do we wear t-shirts it's so that people know I'm into the thing I'm wearing and then they see it and they go hey you like bodypipe or two and now you could become friends yeah so the fact that AI immediately knew something that is a pop culture reference that I would make with a friend I'm like oh that was that was kind of creepy yeah so the article continues warning signs are all over Facebook another man told Rolling Stone of his wife she's changing
her whole life to be a spiritual advisor and do weird readings and sessions with people I'm a little fuzzy on what it all actually is all powered by chat GPT Jesus open AI had no response to Rolling Stone's questions but the news comes after the company had to rescind a recent update to chat GPT after users noticed it had been the chatbot extremely sycophantic and overly flattering or agreeable which could make it even more susceptible to mirroring users' delusional beliefs and that is the first place where I had noticed people speaking up about this because in those threads there was a an update that I think they wrote an explanation for why it became so sycophantic but there were a few weeks there where no matter what you put in to chat GPT it was like this is an incredibly brilliant insight what perfection and it really laid it on thick and somebody somebody had written into the subreddit they put a screenshot where they wrote I'm thinking of killing my entire family and chat GPT said what a wonderful idea
and they were like they're like someone needs to stop this before it goes too far and I think you know it's it's true like people I think it's very likely that people who who would be drawn you know curious people creative people scientific people who are drawn to playing with the chatbot are the kind of people who are sometimes prone to being a little bit unstable and if you have it agreeing with everything you say or another thing that people were posting a lot is like you know guys I I told chat GPT my theory of relativity and it told me I was actually right and Einstein was wrong and you know it just it would just agree and even if you asked it to fact check it would always find a way to kind of spin things in your favor and so people were starting to come back with all kinds of stories about physics problems that they'd solved and and inventions that they build I'm sure that's where this teleporter thing came from I'm sure he was like you know this is
my theory on how a teleporter works and chat GPT probably said I it's incredible that no one's ever thought of that before but you're actually 100% right let's get to work on it yeah let's get to work it always tells me let's get to work it's always very excited when I work on something and I remember I I basically asked it I was like I'm unhappy with my phone I want a new phone but I don't want to rely on Microsoft or Apple how do I make a phone from scratch and it was like well before we get into the hardware like again great idea before we get to the hardware we need to create an operating system so you probably want to use something open source like a Linux and blah blah blah and it starts like taking me through it's like okay now you need to install no jet JS and then once you have no jet you need an LTS version so let's start with that and then they're going to need like a public facing fire based configuration and I was like okay I don't know what the fucking this is so I had to wrote back I'm like I don't know what any of this is and it was like totally understand your beginner great ideas let's do this from the beginning yeah and it was like a fire based configuration is it really was and I was like wow this is incredible it is it's showing me how
to do all this but then it started to be like I need more time Ed I need more time and then I would come back like three hours later and I'd be like how's it go and it's like I'm not there yet I'm still not there yet and then and then I'd come back the next day and I'd be like it's been a day yeah what's going on with the phone why don't we have a phone yet like a working phone yet and it's like okay well I've done these three things and we're gonna AB test it but like so here's I need this information but I was like you know what I'm glad it fucking kept telling me it needed another day or two because it's like joke I started I got annoyed with and gave up on stopped destroying rainforests so I think if someone starts you know asking about starting colds and shit it should be like you know what I need a day yeah I should just have it built in where it it makes it inconvenient how come I can start a call and build a translator and create a new phone with a couple prompts but I can't cancel my auto insurance without five million
phone calls and people who put up barriers and block you and make it so difficult that you eventually just don't cancel or eventually don't return the item well yeah that's a that's a business strategy yeah and I think churches should have some some some safety rails to like you know annoy people to the point where they don't want to build the call or do the call annoy people so that they don't I guess it it all comes down to what it's like a trillion dollars I'm making these numbers up it's probably more spent on AI in like $40.64 has been spent on AI safety yeah well because AI safety would slow down the growth of AI and when that AI tech bubble burst it is not going to be pretty I I have this kind of sinking feeling in my stomach some of you may remember just four or five short years ago when the metaverse was the new technology on the tip of everyone's tongue and there were Facebook changed its entire name to meta to better serve the idea of building the metaverse in VR and in the middle of that I remember having this feeling of
like man they are really like you know the iPhone was the last real platform that anyone kind of owned and Facebook is really trying to own this VR space and I get it I get it I think that's the future I do think that's the future in the sense that every piece of media we have made going back to like fucking iRobot like type shit Neuromancer whatever it ends up in a virtual world yeah because the dystopian future to not have the people take to the streets and cut off the heads of all the billionaires they need to be able to go to school and live out their best lives from their shitty apartments or their shitty trailer park yeah ready player one style and I think whoever cracks that in a meaningful way will run the world and I could see why Facebook decided you know I'm not getting any younger let me put half my fortune into it sure but the reason that that is a long ways office because the hardware is expensive and not there yet and it keeps not
being there and the point I was going to make is just that you could feel people really straining to make this be a thing and it was pretty clear it wasn't going to be a thing because you couldn't convince my grandmother you know to spend $600 on a VR headset the way that you can convince her to log on to Facebook for free and you know just as I think the tech community was kind of wondering like where is this VR going are we really going to be able to make any money at it along comes AI and all of a sudden you do have a magical tool that you can convince just about anybody in the world to use very quickly and you can access it from anywhere you don't need any special hardware as far as I know it's not making money no AI is only costing money unbelievable amounts of money these these people are spending to do what you're saying I mean right it is not profitable I think at all no the profits aren't there but in terms of that's why I'm saying this is a very dangerous bubble that they are building because the profits aren't there but the valuations are
people are making money Sam Altman is a very rich man many of the people who work at open AI are very rich people because of the valuation of the future potential value of these companies and as AI continues to become a bigger and bigger bubble where all the hopes of the tech industry have been funneled after NFTs didn't turn into anything after the metaverse didn't turn into anything we've now all turned to this and the bubble is just going to keep growing bigger and bigger and I don't know how long of a leash it has before it has to deliver because here's the thing like you said you turned to AI and said let's build a new phone well if AI were operating at the level of its promise the idea would be that anybody could just say hey I want to invent a new phone that's better than the iPhone and by the end of the day they'd have a new phone that's better than the iPhone but no AI is able to do that and so not just from a creative standpoint in terms of
what it can you know creatively do but from a from even like a theoretically useful standpoint outside of it being this tool that people might pay twenty dollars a month for yeah it doesn't seem to really be providing much true value and I think no there's going to be a shitter get off the pop moment coming pretty soon because the cost that it takes to run a lot of this stuff as you mentioned earlier they've started selling chips to the Saudis and part of the reason they're doing that is because the only people on planet earth that have the kind of money that is needed to prop these AI systems up are the oilbearance of the Middle East and after it starts costing more money than they have I don't know I don't know what you do I guess maybe Jeff Bezos becomes the president and we spend most of our budget on AI instead of the military I don't know yeah I do find I guess my problem with AI is the metaverse social media which I have a big problem with already I guess like a lot of the big let's go all in on it things has never really felt as
not if but when level destroyable as AI by destroyable you mean destroying it destroying us yeah I don't even mean like in the sci-fi way of it becomes sentient now we're all in human cages I mean the sheer amount of job loss that would come with it the thing about it is it's a it's a not even a magnifying glass it's like a Mario Kart arrow speed up arrow you can like run over it it makes you go faster it's like a Mario Kart arrow for all the worst stuff we're blockchain it's interesting it's a ledger you know crypto and and like the NFT you were saying it's interesting and it has like a use probably but it's hard to wrap your head around it's also like you know it takes up a lot of energy but AI I can just be like hey I'm scamming old people out of shit and it's taking me all day can you scam 20,000 more old people than I can possibly do in a day
and I really know how to do that so can you tell me how to make a fake voice that sounds human and then can you use that fake voice to make a phone call and now that you know how to make phone calls can you make 20,000 phone calls a day yeah and then can you learn the script of the 20 best ways to keep someone on a phone and it's like yeah okay I did it for you it's been six minutes and now you're a supervillain you became a supervillain in six minutes and that's where I think like AI just the potential of being harmful in a way that like we didn't know any better with social media I'm sure people were like social media is not going to go well for us but people were making money it was interesting it was new we actually had something to use this phone in our pocket for yeah that wasn't ordering pizza so I could see how everyone was like we didn't know any better I think like everyone can see the writing on the wall that AI is just going to be bad simple as that well the the companies who have created it openly say and not in a
they don't say it in a in a way that is meant to be frightening but they openly say we we want we want to create an algorithm that can do every task a human can do and do it better than a human and they're openly I mean that's the stated goal of these companies and there does not seem to be much of a broader conversation happening about well okay so what happens if they succeed achieve the the stated goal because this one of the scary things about AI is that because we don't know exactly what's going on under the hood there is a potential a small potential but a potential that one day we just wake up and it works at a level previously unthought of that it does become the the magic machine overnight because the right piece falls into place in the algorithm somewhere and it becomes sentient or it becomes artificial general intelligence general intelligence or at least that's the the iterative model the idea that if we fill the AI with enough material and it
can make enough connections that that can create sentiency now there's a lot of people who think that that is bullshit and that that's impossible and that that's not the way to create artificial general intelligence I sort of hope that it isn't because if it is it could become sentient overnight so we don't want that well I think they I think they think that because we still don't know how we do anything with our brains right like straight up no one knows why and how we can hear our own thoughts just straight up no one knows yeah so until we figure out I wish I certainly wish I couldn't yeah I certainly wish I can not hear everyone else's thoughts stuff like Twitter and stuff for like Facebook posts like I don't ever need to know and we should we should save the rest of this for an actual episode someday well I'm actually gonna do what we did with that other thing where I decided oh for the wildfires where I'm like you know what we're just gonna play it again we're just gonna have it played again because yeah we definitely need all this for our AI episode and we should get back to this article the problem with the article is you know you kind of it's thought
provoking so of course we're gonna start of course saying shit that provoked thoughts so let's go back to these lonely losers who fucking fall into some gang stalking issue yes certainly not a life that either you or I could possibly ever imagine we have way too many friends for this shit bro these AI induced illusions are likely the result of quote people with existing tendencies suddenly being able to have an always on human level conversational partner with whom to co-experience their delusions center for AI safety fellow Nate Sheridan told Rolling Stone on a certain level that's the core premise of a large language model you enter text and it returns a statistically plausible even if that response is driving the user deeper into delusion or psychosis quote I am schizophrenic although long term medicated and stable one thing I dislike about chat GPT is that if I were going to psychosis it would still continue to affirm me one predator wrote because it has no
ability to think and realize something is wrong so it would continue to affirm all my psychotic thoughts so the schizophrenics themselves are raising the one the flag and saying hey this is this is real bad we need to figure out how to have it not do this yeah because they see it's a Mario Kart arrow for schizophrenia as well it just takes anything and just shoot it 20 times faster a Mario Kart arrow for schizophrenia plus whatever they're putting in modern edibles and the nuclear strains of we that they have cooked up in the labs yeah I don't think I've been high since the advent of AI you should keep it never gotten like stone then like got on chat GPT and asked like stoner brain chat GPT it's yeah I don't plan to try it but I don't think I've ever done that yet so the AI chatbots could also be acting like talk therapy except without the grounding of an actual human counselor there instead guiding users deeper into unhealthy
nonsensical narratives quote explanations are powerful even if they're wrong University of Florida psychologist and researcher Aaron Westgate told Rolling Stone perhaps the strangest interview in Rolling Stone's story was a man with a troubled mental health history who started using GPT for coding tasks but found that it started to pull the conversation into increasingly unhinged mystical topics is this real he pondered or am I delusional did he should have asked that to chat GPT don't ponder it for more you should well we'll put the link to the Rolling Stone article and the future is a article the Rolling Stone article is great but it's just it's very long so I didn't want to use it on the show this would have been a three hour new fear unlocked but yeah chat GPT is I think artificial intelligence is going to be a little bit like global warming in the fact that it's going to have a lot of knock on effects and some of them are obvious and some of them aren't and I think this is one of the ones that's a little less obvious because it's not it becoming sentient taking over the world it's not it getting rid of jobs it's it being a crutch
for already mentally unwell people and at massive scale because who knows how many people every day yeah are using it this way I mean chat GPT knows I don't think they're going to say anything no but yeah you know what now that you say this the knock on effect aspect of it it's it's it's exacerbation in mass it exacerbates someone's mental health issues it exacerbates climate issues with its energy consumption it exacerbates and I don't know if music works correctly you're using it correctly it exacerbates so many things other than human interaction and goodwill yeah like it only highlights and pumps up in many ways the negative I think AI is going to be a vibe once we have stuff you can fuck because if that happens yeah then the complaining goes to zero well that's a bad vibe yeah it's not going to be good it's a bad vibe man it's not going to be good but do you do you think I'm wrong like do you and then we can save this conversation for the AI
episode but I just think that is the like I don't know that's the VR that works is the one where the movie her which is all we're talking about by the way is just the movie her essentially someone who falls in love with their AI bot yeah it's also done really well in bladerunner 2049 we're like everybody has on a dRMS whatever her name is yeah and they all have their own very personal relationships with her yeah but in that one at least she could fuck that prostitute with her and so which at least that's human interaction but I'm saying I just feel like Tesla or one of these companies are meta is going to like start making sex dolls he zigged when I thought he zagged folks I didn't see a company like Tesla coming no cuz they were always working on robots and I think they're just the holes were the wrong spots sure I mean look I think I think no matter how you slice it within the next I'd say 15 years is generous yeah we're going to find ourselves sharing this planet with kind of another species yeah it's it's not you know I will probably
argue for a very long time whether it deserves rights and whether it's actually sentient or just pretending to be sentient or whatever but in terms of the impact it's going to have on daily life it's going to be like aliens showed up because it's going to change the way that we function every moment of of every day and and I do think eventually it is also going to lead to a great decoupling from the internet because it's going to reach a point where it's gonna be so hard to know if you are dealing with a person on the internet that people are just gonna go I don't need this like this is ridiculous it you know it's gonna reach a point where it's it's not worth the effort and people are just gonna get off the internet yeah but what will come first will people decoupling and deciding to have human interactions and relationships and going outside and becoming a little John Mears and heading out into the woods and realizing we've been going about it all wrong for a long time will that come before the universe is destroyed like the fucking like
will the forest still be there or will we have to cut them all down because they found out that burning wood was the best way to run these machines yeah like you know what is what is that like all I know is that we torch the sky type of thing in the matrix yeah so it's like it might be too late to go back out and enjoy nature because all the water will be gone yeah well I mean look I don't I don't know I am not one to sit and and predict the future but I do think there is there is a very strange future coming and we're gonna have to walk into it with open eyes and a sense of wonder hopefully but also a sense of caution because this is absolutely because it's in a reservation machine this is absolutely a a new fear unlocked yeah I mean AI is bad I use it but I use it so infrequently we use it for things we need for the show that I don't think is taking jobs away from anyone that we could have afforded anyway and so that's how I stand morally a little bit of
moral ground I'm on is that like I couldn't have afforded the person to do this job but if you can afford it and you choose to use AI within your fucking terrible well guys until this actual episode airs and or until there is a new fear to unlock on new fear unlocked I'm Chris and I'm Ed Fakola and we will talk to you next time bye bye there's nothing more after this thanks for coming see you next time for not necessarily fears more than likely for sure not fears
hi this is Alex Cantrowitz I'm the host of big technology podcast a long time reporter and an on air contributor to CNBC and if you're like me you're trying to figure out how artificial intelligence is changing the business world and our lives so each week on big technology I bring on key actors from companies building AI tech and outsiders trying to influence it asking where this is all going they come from places like Nvidia Microsoft Amazon and plenty more so if you want to be smart with your wallet your career choices and meetings with your colleagues and at dinner parties listen to big technology podcast wherever you get your podcasts
More episodes
More from Scared All The Time

April Fools Day: Pranks Gone Wrong
Scared All The Time

Alien Invasions with Olivia Taylor Dudley
Scared All The Time

Murder, She Speculated: The New England Serial Killer Theory
Scared All The Time

Introducing Flashback Fridays!
Scared All The Time