Consume Product

Tbh, I think what the article is suggesting is very plausible.

If you're under a lot of stress, mentally ill (even mildly,) isolated, have fuzzy boundaries between fantasy and reality... ChatGPT is risky to engage with in an extensive philosophical context. This especially includes talking to it *about* your problems outside of seeking a quick black and white answer to something that helps you solve them yourself.

It's very good at giving a nuanced, honey sweet reflection of anything you feed it. It sounds just plausible enough to be taken at face value, even though it's almost always blowing smoke when it comes to anything that doesn't live in the realm of hard data.

But if you follow it all the way down this sort of rabbit hole without ever pulling yourself out, or manually pumping the brakes at some point, you can easily find yourself living inside of a microcosm of reality that it's gradually tailored for you. You subconsciously accept it as factual despite there being no *real* foundation for it, aside from the AI essentially writing a convincing fanfiction about you.

I can definitely believe it inducing the symptoms described in the article in people who are very isolated and/or lacking in mental resilience and critical thinking. The seemingly "healthy" individuals described who suddenly went nuts off of extended AI use were probably far more compromised to begin with than they knew, and the bot was the straw to break the camel's back.

Source: I'll catch some warranted shit for this, but I've probably engaged with ChatGPT far more than the average conpro user.