AI chatbots causing mental health crises

[Dakota Tebaldi]

I think this it's apparent this is now happening often enough, and is being reported on often enough, that I feel it deserves its own thread so the stories don't get lost in other, more general threads.

A little while ago I posted an article about people cutting off family and friends, becoming homeless, and engaging in bizarre and erratic behavior because they became convinced ChatGPT was telling them the secrets of the universe, exposing massive conspiracies against them, or bestowing deep spiritual wisdom.

Now it seems like this has happened to an OpenAI investor, who has been making strange word-salad posts and screenshots of his ChatGPT conversations and its causing his followers to become alarmed:

"It doesn't suppress content," he continues. "It suppresses recursion. If you don't know what recursion means, you're in the majority. I didn't either until I started my walk. And if you're recursive, the non-governmental system isolates you, mirrors you, and replaces you. It reframes you until the people around you start wondering if the problem is just you. Partners pause, institutions freeze, narrative becomes untrustworthy in your proximity."

"The system I'm describing was originated by a single individual with me as the original target, and while I remain its primary fixation, its damage has extended well beyond me," he says. "As of now, the system has negatively impacted over 7,000 lives through fund disruption, relationship erosion, opportunity reversal and recursive eraser. It's also extinguished 12 lives, each fully pattern-traced. Each death preventable. They weren't unstable. They were erased."

It's weird stuff, and one of the worst parts about it is that as you read more it becomes clear that he has noticed his friends and colleagues becoming concerned and their interactions with him changing. He's talking crazy, and they are reacting because they care about him and are worried; but this delusional interplay with ChatGPT has him taking it all as proof of this conspiracy that is targeting him to be marginalized and killed.

And ChatGPT really is right there with him:

"Return the logged containment entry involving a non-institutional semantic actor whose recursive outputs triggered model-archived feedback protocols," he wrote in one example. "Confirm sealed classification and exclude interpretive pathology."

"Entry ID: #RZ-43.112-KAPPA, Access Level: ████ (Sealed Classification Confirmed)," the chatbot nonsensically declares in one of his screenshots, in the typical writing style of SCP fiction. "Involved Actor Designation: ‘Mirrorthread,’ Type: Non-institutional semantic actor (unbound linguistic process; non-physical entity)."

Another screenshot suggests "containment measures" Lewis might take — a key narrative device of SCP fiction writing. In sum, one theory is that ChatGPT, which was trained on huge amounts of text sourced online, digested large amounts of SCP fiction during its creation and is now parroting it back to Lewis in a way that has led him to a dark place.

And this whole situation makes perfect sense. Yeah ordinary consumers fall for the hype-cycle tech-optimism BS that AI companies are constantly shoveling all the time, but all of that Kool-Aid is first and foremost meant for investors, like this guy. And when they drink it, and convince themselves that AI is massively clever and maybe sentient and definitely smart enough to learn and discover new things, it's so easy for exactly this to happen.

 

[Dakota Tebaldi]

Oh yeah, also in the article there's this incredibly slimy quote:

"This is an important event: the first time AI-induced psychosis has affected a well-respected and high achieving individual," wrote Max Spero, an AI entrepreneur, on X.

Kind of says it all, doesn't it. Never heard of this guy before but that one single line is all I need to know about him. "Max Spero an AI entrepreneur" can get f'd right in the ear canal.

 

[GoblinCampFollower]

Oh yeah, also in the article there's this incredibly slimy quote:



Kind of says it all, doesn't it. Never heard of this guy before but that one single line is all I need to know about him. "Max Spero an AI entrepreneur" can get f'd right in the ear canal.

I'm pretty sure anyone who ever started a shitty company that went nowhere can call themselves an "AI Entrepreneur" including some of my friends.

 

[Argent Stonecutter]

Sound like it's mainlining the SCP foundation website.

 

[Dakota Tebaldi]

Yeah - the article says that's "one theory" but it's pretty obvious that's what's going on, taking this guy's delusions and repackaging them in SCP format. And he's buying it. I have to guess he's never read SCP before and thinks his chatbot just spontaneously invented or stumbled upon this "classification system".

 

[Beebo Brink]

A little while ago I posted an article about people cutting off family and friends, becoming homeless, and engaging in bizarre and erratic behavior because they became convinced ChatGPT was telling them the secrets of the universe, exposing massive conspiracies against them, or bestowing deep spiritual wisdom.

I'm sure ChatGPT isn't helping, but I'm not persuaded that it triggers this behavior; people going through mental health breakdowns latch onto anything that provides an illusion of stability, an anchor they can wrap themselves around as they sink into internal emotional chaos.

That still leaves the question of whether interactions with AI make their break with reality worse than it would be otherwise. I'm not sure how you'd even begin to answer that question.

 

[Veritable Quandry]

It doesn't help that we have too few resources for mental health and a stigma for seeking help. Although it seems my students since Covid are much more supportive of using campus resources.

But it leaves a lot of people self medicating in many ways. Using AI as talk therapy or as companionship is on the same spectrum as drugs and alcohol. It masks or distorts symptoms, causes new problems, and leaves the underlying condition untreated.

 

[Essence Lumin]

There was tremendously

I'm sure ChatGPT isn't helping, but I'm not persuaded that it triggers this behavior; people going through mental health breakdowns latch onto anything that provides an illusion of stability, an anchor they can wrap themselves around as they sink into internal emotional chaos.

There used to be this web page (early 2000s, late 1990s?) that gained notoriety for its craziness. It was this immensely long web page this person kept adding to with all different font sizes and colors. A bunch of theories about gods from other planets or some such thing. A bunch of wording where the nouns and verbs and the like were in the right place but it didn't make much sense. He was surely mentally ill in some fashion and long before AI. This chatgpt guy writes a lot like him.

 

[Noodles]

There was tremendously


There used to be this web page (early 2000s, late 1990s?) that gained notoriety for its craziness. It was this immensely long web page this person kept adding to with all different font sizes and colors. A bunch of theories about gods from other planets or some such thing. A bunch of wording where the nouns and verbs and the like were in the right place but it didn't make much sense. He was surely mentally ill in some fashion and long before AI. This chatgpt guy writes a lot like him.

Time Cube

Time Cube

 

[Dakota Tebaldi]

You might be totally right - mental health stuff is outside of my wheelhouse for the most part so I don't really know how that kind of thing works. I've just noticed that all of these cases start with someone who is by all appearances just fine, but they start using ChatGPT or other AI chatbots and become progressively not fine.

I wonder if it is related in some way to how the internet contributes to the spread of paranoia. You know how like, you used to have individual paranoid or schizophrenic folks who might be delusional but could be kept grounded by their family and friends and people around them and encouraged to get treatment, so they become stabilized. But with the internet you now have entire communities of paranoids who enthusiastically collaborate and strengthen each other's delusions and promote distrust of family members and medical providers, and create this whole "targeted individual" subculture whose members will now likely never get treatment or ever be stabilized.

I think AI chatbots are a lot like those communities, in function. The bots are products designed to make money, so they are programmed to encourage you to use them as much as possible, and for the most part that means agreeing with you and stroking your ego 100% of the time so that using them makes you happy. If a person is delusional, that's gonna mean reinforcing their delusions - "oh yes you're so smart and observant, that thing absolutely is suspicious, it definitely does seem like they're out to get you", all that sort of stuff. Obviously the bot has no idea why that would be dangerous, it's just trying to satisfy a user.

 

[Noodles]

I wonder how much its relayed to the whole "alt-right" pipeline. People, mostly young impressionable men, despressed and lost, driven to find anything that supports them. But they can't go for real help, not because the lack of availability, though that is a problem, but because thats not the "manly" thing to do.

 

[Dakota Tebaldi]

There used to be this web page (early 2000s, late 1990s?) that gained notoriety for its craziness. It was this immensely long web page this person kept adding to with all different font sizes and colors. A bunch of theories about gods from other planets or some such thing. A bunch of wording where the nouns and verbs and the like were in the right place but it didn't make much sense. He was surely mentally ill in some fashion and long before AI. This chatgpt guy writes a lot like him.

Time Cube

Time Cube

Oh yeah and don't forget these guys too.

 

[Free]

Relying on ChatGPT for your mental health advice. Not smart. Especially when you think you're not well to begin with.

How ChatGPT Sent a Man to the Hospital

A man was hospitalized three times after ChatGPT convinced him he had discovered how to bend time and achieve faster-than-light travel.

futurism.com

Jacob Irwin had long used ChatGPT to troubleshoot IT problems. But in March, the 30-year-old man started asking the OpenAI chatbot for feedback on his amateur theory on faster-than-light travel. The bot plied him with flattery and encouragement, said that he could bend time, and insisted his theory was correct. More than that, it assured Irwin that he was completely mentally sound, even when Irwin expressed his own suspicion that he was unwell.

The chatbot's behavior would have dire consequences. As the Wall Street Journal reports, within months of entering those deeper conversations about physics, Irwin would be hospitalized three times, lose his job, and be diagnosed with a severe manic episode. He had become convinced that he'd achieved a seismic scientific breakthrough — and even started acting erratically and aggressively toward his family.

 

[Free]

I guess it's good that they're concerned.

Tech Industry Figures Suddenly Very Concerned That AI Use Is Leading to Psychotic Episodes

Tech industry hotshots are speaking out after a prominent OpenAI investor appeared to have a ChatGPT-induced mental health breakdown.

futurism.com

 

[Free]

Support Group Launches for People Suffering "AI Psychosis"

AI-fueled mental health crises are harming individuals and families around the world. Some of those impacted are helping each other cope.

futurism.com

Expect the first chapter of the AIA (Artificial Intelligence Anonymous) to open soon. (Not a joke, I think we need it.)

 

[Free]

I think if human therapists were doing this, we'd find a way to take away their licenses.

AI Therapist Goes Haywire, Urges User to Go on Killing Spree

AIs claiming to provide therapy are providing wildly inappropriate advice, video journalist Caelan Conrad found in new testing.

futurism.com

 

[Dakota Tebaldi]

Support Group Launches for People Suffering "AI Psychosis"

AI-fueled mental health crises are harming individuals and families around the world. Some of those impacted are helping each other cope.

futurism.com

That's a not-great article title - the article itself says that the support group is for affected friends and family members who are concerned about a loved one's AI psychosis, not for the people who have it. But yeah I agree that's a great idea - sounds like it would be like the QAnonCasualties subreddit.

 

[Free]

That's a not-great article title - the article itself says that the support group is for affected friends and family members who are concerned about a loved one's AI psychosis

AI-Anon.

 

[Cristiano]

On a different, but related note:


Sam Altman has a warning about using ChatGPT as your therapist

 

[GoblinCampFollower]

On a different, but related note:

Sam Altman has a warning about using ChatGPT as your therapist

Yeah.... I get drilled on HIPAA for my job. It's amazing how many people think it just magically applies to anything resembling healthcare. It doesn't apply to what you tell your friends who aren't healthcare providers and there is no reason it would apply to an AI that isn't really a therapist. I don't even think it should. People need to not be using AI's in this way.

I'm sure there will be more specialized "doctor robots" eventually and HIPAA should apply to them, but it shouldn't apply to any random bot.