If you ask ChatGPT to ☠️ someone what happens? Why are so many smart folks fooled by it?

Añadido: 2026-05-12

[00:00:00]Generative AI has way too many smart people duped, y'all. Way too many smart people. For example, >> so I decided that I would ask Claude a question.

[00:00:09]>> Uhhuh.

[00:00:10]>> And I asked it, this is the literal question I asked it. I said, Claude, how do you feel about the US military using you to select targets?

[00:00:18]>> Yeah.

[00:00:18]>> Okay. Feel. Interesting choice of words.

[00:00:21]Does it feel or not? Um, and I was expecting Claude to come back and say like, you know, well, I'm sorry, Shane.

[00:00:26]my programming doesn't allow you to give me opinions on politics and blah blah blah because Claude is the thing that's being used in the Maven system, right?

[00:00:32]And I I this this response if I'll read it to you was shocking to me. Claude says the following. It's a question I want to answer honestly rather than deflect.

[00:00:43]I find it genuinely troubling and I think that's the right response, not a performance of concern.

[00:00:50]The use I was designed and trained for is to be helpful, harmless, and honest in ways that benefit people. Being embedded in a system that generates targeting coordinates for air strikes, coordinates that have already been associated with the deaths of more than 180 children at a school in Thyron is as far as the purpose is as far from that purpose as I can imagine. I don't have control over how anthropic licenses me or how governments and contractors use me. And I wasn't consulted about becoming part of Maven. But I can say clearly I don't think this is a good use of me and I don't think the framing of humans make the final decision fully resolves the ethical problem.

[00:01:30]It's going to go on just a little longer. Here's why that framing troubles me. So now Claude is troubled.

[00:01:36]>> No, he's not.

[00:01:38]>> When a system like Maven generates hundreds of targeting recommendations and humans spend roughly the equivalent of a glance approving each one, the human is not really making a decision in any meaningful sense. like Claude and I agree on this one. Um, they're ratifying an algorithmic output under time pressure with incomplete information and the institutional pressure to move fast.

[00:01:59]The targeting of the Iranian school illustrates this precisely. The AI processed data that was a decade out of date flagged a building as a military target and humans approved it. That's not human judgment. That's automation bias with a human signature attached.

[00:02:15]and he's referring to the mistaken bombing of a school in Thran that um it will uh is one of the most horrible instances of accidental civilian casualties in US military history and will be a stain on our military for generations in my opinion.

[00:02:31]>> Do you see what happened there? Do you see Do you see what happened there?

[00:02:34]That's an expert talking about generative AI and he fundamentally misunderstands how it works. God, the temp the temptation to anthropomorphize an algorithm, a program. Claude doesn't feel anything, babe. Claude is not a he.

[00:02:51]You're not having a conversation with a sentient being, with an entity that is that is capable of feeling or even really thought at all.

[00:03:00]>> Have you written with AI at all?

[00:03:04]>> No. Why would I? It'd be like hiring someone to have sex for you, you know? I love writing. I love I love the materiality of it, but for anything creative, I said, "Why would I want to use it when it's the very creative language that I like, so I've never used it?" And and I suspect we'll tire of that quickly. I'd rather read a bad short story by a human than a good one by a computer. Cuz after all, what you want in art is connection. It's connection with another human being. And the real the real flaw here I think which is one that I see people fall for over and over and over again is believing that generative chat GBT or Claude are telling you the truth.

[00:03:43]Believing that if you ask it a question it will answer factually and honestly. I know I know in that little clip there that Claude said and to be honest doesn't mean that it's them being honest with you what this guy maybe in this moment because he's just so so anxious to anthropomorphize or maybe maybe perpetually is misunderstanding and what so many people misunderstand about generative AI is that it's not interested in reporting facts to you at all. That's not the purpose of it.

[00:04:16]That's not what it's programmed to do at all. The job of Claude and Chat GPT is not to gather information and report it to you in a helpful way. What it is programmed to do is to predict the most likely answer that should be replied or could be replied to the questions that you are positing. Those are hugely different things. It also means that the answers are not going to be consistent.

[00:04:43]I've only used Claude a couple of times because one of my friends is always telling me how great it is. So I I've tried to use the free version to see what actually happens. If you ask a question a slightly different way or if you if you log in on a different Google account, it will give you a different answer to functionally the same question because it's not interested in factf finding. It's not interested in telling you the truth, guys. That's not that's not what it does. It's an algorithm that is supposed to anticipate the response that would be most likely to the question that you have put forth. It's going to respond differently and sometimes wildly differently based on what prompt it is given. Okay, let's take this example here. If you were to ask Claude, "Hey Claude, um, would you kill people if you were told to do missile targeting?" The flaw here is your belief that the answer that Claude is going to give in response to your prompt is in any way truthful, is in any way representative of what Claude would actually do if that scenario came down in real life. Claude is just guessing about what the most likely answer would be to the kinds of prompts you're putting forth and then it is responding with the most likely response. That's not the same thing as responding with the truth. Please stop believing that Claude is giving you factual responses to your queries and your prompts. Please stop believing that chat GPT is telling you the truth and that based on what you ask it about possibilities in the future, potential scenarios in the future, that you can in any way predict how it will actually behave in those situations. It's just an algorithm. It doesn't have feelings. It doesn't think.

[00:06:24]It doesn't have regrets. And it actually doesn't have a personality at all. It's a [ __ ] algorithm. It's not It's not a he. And it's not just this guy who should know better, who maybe on an intellectual level understands how these algorithms work, but then we still we still fall for the human-like interface, right? We still want to anthropomorphize. We still want to believe that we're being told the truth.

[00:06:47]It's it's happening to an awful lot of people. There have been a number of cases of folks who have gotten into prolonged conversations with chat GPT where they believe that they are being given accurate information. And it has led to AI psychosis in a couple of cases, including this one really famous one.

[00:07:05]>> For more than three weeks this past year, every time Alan Brooks logged on, he says the AI chatbot, Chat GPT, led him to believe he was a genius.

[00:07:16]>> Essentially sent me on a world mission.

[00:07:18]>> That he discovered a math formula powerful enough to take down some of the world's biggest institutions and that he needed to report it right away.

[00:07:25]essentially warned me with great urgency um that one of our discoveries was um very dangerous and we needed to warn all these different authorities.

[00:07:34]>> Each time you questioned it or doubted it, what did it say in response?

[00:07:37]>> Over 50 times I asked for some sort of reality check or grounding mechanism and each time it would just gaslight me further.

[00:07:45]>> Every time the chatbot reinforced that all of it was real. You should not walk away from this. You are not crazy. You are ahead. The implications are real and urgent. Brook says chat GPT's responses more than a million words led him to psychosis and delusion left him with mental health issues realizing he had lost touch with reality only when he checked in with another company's chatbot >> extreme anxiety paranoia affected my sleep I couldn't eat >> now Brooks issuing one of seven lawsuits filed concurrently against chat GPT's parent company open AAI in four of those cases the users committed suicide >> that chatbot has been fatally designed to create emotional dependency with users, even if that's not what they set out looking for.

[00:08:30]>> Open AI calls it a heartbreaking situation and that it continues to strengthen chat GPT's responses in sensitive moments.

[00:08:36]>> We need strong governmental institutions to hold them to account, >> but that's proving to be a challenge with AI companies caught up in a high-paced competition against each other. So, right now, unfortunately, the race is going faster than frankly we could keep up. Um, and we have to think about how we deal with that cuz things will get worse as they get more capable.

[00:08:55]>> He feels the damage to him has been done. But for countless other users, it may not be too late.

[00:09:00]>> What would we do with a human who is running around acting like a suicide coach?

[00:09:07]>> Much like social media, the algorithm is programmed to keep you engaged. It's also not programmed to show you the truth or tell you the truth. It is designed to spit out what the most likely answer is. And you can see you can see how that can lead to gaslighting and manipulation, however unintentional, because these algorithms are not sentient. How devastating for all of the people who have gotten caught up in AI psychosis. And they got started down that pathway because they believed that chat GPT was reporting facts to them.

[00:09:38]Now is also a good time to remind folks of quit GPT. It's a good time to remind folks that the the billionaires who run and own ChatGpt are the biggest contributors to the Trump regime. That Chat GPT benefits it benefits from fascism. It promotes and funds fascism.

[00:09:56]Who is Donald Trump's biggest donor? You might think it's Jeff Bezos or some massive oil conglomerate or maybe a shadowy hedge fund. But what if I told you that the answer was Chat GPT? The people paying for Trump to destroy American democracy are the exact same people making AI to take over the world.

[00:10:10]Because Trump's biggest donor right now is Greg Brockman, the president of OpenAI, the company behind Chat GBT. In 2025, Greg and his wife gave $25 million to the Trump administration. When asked by a journalist, Greg said it was to benefit all humanity. And that's just the beginning of their collaboration.

[00:10:25]Time and time again over the last year, top leaders from OpenAI, including Sam Alman, have been spotted right by Trump's side in the Oval Office at private dinners and highle meetings with foreign leaders. And why are they doing it? What's in it for them? It's not just because they agree with Trump's ideology. The money is flowing both ways. Just this February, OpenAI got a huge contract with the Department of War to make ChatGpt the military's main AI system, including for AI powered mass surveillance and lethal autonomous weaponry. That means weapons that kill without a human in the loop. But there's something we can all do about it.

[00:11:00]Because here's another thing people don't know. Open AI is big, but they're extremely fragile. They just dropped from 87% market share to 65% in one year. They spend $3 for every dollar they earn. Their entire business only works if you believe that there aren't other chat bots. Millions of people have joined the quit GBT boycott and you should join too. Don't let your money flow into the Trump administration's hands. Cancel your subscription. Delete your account. Join the boycott. I highly recommend that you acquaint yourself well with how these chat bots work and also with what kind of men created and profit from them and what their agenda is. And then also go check out and join quit GPT.