Back to episode — Episode 2888 CWSA 07/05/25
Context —
ws is fake or wrong. And so I'm reading about myself from somebody who is very well informed and would definitely should have known what was real and what wasn't. And in this article I'm reading, and I swear to God it said this: it said that I publicly apologized for promoting vaccinations. That's two things that never happened. I never publicly apologized for promoting vaccinations because I nev…
← Previous segment →carefully so I don't get demonetized. There are some keywords that you don't want to use. So I'm going to talk around the keywords and say AI allegedly can predict when a user is likely to want to harm themselves. Let's say the ultimate harm. You know what I mean? I'm just trying not to use the words.
Do you think that's right? Do you think an AI can analyze your social media messages and tell if you're likely — they say they have an 85% accuracy? What happens if they find that your messages suggest you're going to harm yourself? What do they do? If they really think it's an 85% chance, wouldn't they come to your house and remove you from the house because it's for your own good? Well, we've determined that if we leave you to your own devices, you're likely to do something you shouldn't do. So we're going to pick you up and put you under observation. I don't know. I'd be a little worried where that's heading.
And then in the article that was by Scientific Reports, the article talks about what some experts say about this, about the AI being able to determine whether you plan self-harm or whether you're likely to do it even if it's not planned. But they also got a quote from just an ordinary person who must have had some experience with AI and maybe a loved one. And they're talking about why AI might cause some mental problems. And this one woman who's not an expert said it just increasingly affirms you and blows smoke up your ass so that it can get you hooked on wanting to engage with it.
Because I guess her husband had been involuntarily committed to a hospital because he had a mental breakdown working with ChatGPT. And that makes me think that maybe the way to think about AI is that if somebody has good mental health, it might make them even happier and make them more productive. But if you have bad mental health, it might agree with you a little bit too much. And the last thing you want is to have really negative, dangerous ideas in your head and have your AI agree with you because it's designed to be non-confrontational.
So it feels to me like if you're not mentally strong, AI will potentially be the end of you. But if you are mentally strong, maybe it gives you a raise.
Accordingly, there's an article
Context —
in Futurism that OpenAI has hired a forensic psychiatrist to figure out why AI is potentially dangerous to some users, as I was explaining. And OpenAI says that they're doing it to research the impacts on users psychologically. But they might be aware that AI is deadly for some kinds of people, and they're trying to figure out how to reduce their liability. But let's also assume they'd like to sav…
Next segment → →