The lawsuit, which was filed at the San Francisco County Superior Court on 21 July, details how Winter was a regular ChatGPT user and came to rely it on it for medical advice, resisting efforts from family members and friends to make him go to hospital.
This dipshit is why we have so many safety labels on things
When I was being encouraged to use Copilot at work, MS made it clear that their product was intended for "entertainment purposes only." I was sooooooooooo entertained.
Imagine how entertaining it would be to get medical advice from an LLM instead of advice on how to do the job I did just fine for 10 years before LLMs were invented. Very, very entertaining.
Winters' lawyers alleged the platform "repeatedly and insistently misdiagnosed" him while drawing on previous discussions about religion to assure him he would recover.
I forgot that it "learns" now, I haven't tried it in a long while š¤¦
"American man" is all the explanation you need here. If you have to pay for doctors, of course people who can't pay are going to turn to a free alternative that seems to offer some hope. That's one of the reasons why you don't have to pay to see a doctor in civilized countries.
I believe the commenter meant that the price of medical care in the USA push people to extensively rely on AI (or online communities btw), not that Americans are especially gullible or prone to AI psychosis.
I clicked on the article to make sure: he used regular ChatGPT, not some specifically medical AI chatbot.
It's like with Fox "News": it's not news, it's entertainment. Taking it seriously is kinda dumb*. But at the same time, that's no reason to just let it fester and destroy people.
It's actually worse than Fox News. The companies are pushing all this not only like it actually works, but like it's the second coming of Christ. Which is criminally insane.
Part of me wanted to blithely reply that "nothing of value was lost"; However, it's undeniable that this is a societal problem and this person is a victim every bit as much as someone who gets injured in a Tesla because it decides to run a red light and slam into a tree after a bad software update.
These products are built to be manipulative and harmful in the long-term, and by rights, should be regulated into oblivion. The LLM companies are just in the process of calibrating and tweaking the product so it kills more like cigarettes and less like interacting with a Terminator, and for that to happen, a few people have to die. They've almost certainly calculated out the potential litigation costs and accepted the risk.
And most of us can't avoid these products entirely because our job requires us to develop a basic proficiency in their use.
They do, but you can't say "this isn't medical advice" and then turn around and give medical advice. Especially to the point of discouraging seeing a doctor.
22 Comments
Eyekaytee@aussie.zone · 47 pts · 10d
This dipshit is why we have so many safety labels on things
lemmy_outta_here@lemmy.world · 23 pts · 10d
When I was being encouraged to use Copilot at work, MS made it clear that their product was intended for "entertainment purposes only." I was sooooooooooo entertained.
Imagine how entertaining it would be to get medical advice from an LLM instead of advice on how to do the job I did just fine for 10 years before LLMs were invented. Very, very entertaining.
otter@lemmy.ca · 18 pts · 10d
I forgot that it "learns" now, I haven't tried it in a long while š¤¦
panda_abyss@lemmy.ca · 7 pts · 10d
By learning it just creates a list of āfactsā it extracts from the conversation.
Thereās no updating model weights.
(Though I suspect you know that due to your use of quotes this is not for posterity)
Speculater@lemmy.world · 4 pts · 10d
It doesn't "learn" people are too stupid to start a new thread. That brings it back to default.
Zarobi@aussie.zone · 6 pts · 10d
It didn't used to. There was the whole "memories" thing that it kept between conversations. I hated it
Eyekaytee@aussie.zone · 4 pts · 10d
Duno why you're being downvoted, memories are a thing that all cloud AI providers have
Zarobi@aussie.zone · 5 pts · 10d
I remove my own default upvote from my comments because I'm a weirdo
Eyekaytee@aussie.zone · 5 pts · 10d
fair play š
nyan@lemmy.cafe · 17 pts · 10d
"American man" is all the explanation you need here. If you have to pay for doctors, of course people who can't pay are going to turn to a free alternative that seems to offer some hope. That's one of the reasons why you don't have to pay to see a doctor in civilized countries.
FlashMobOfOne@lemmy.world · 3 pts · 9d
It's really not. AI Psychosis is not exclusive to the US, nor is LLM use. These products are every bit as harmful to everyone else outside the US.
leriotdelac@lemmy.zip · 1 pts · 7d
I believe the commenter meant that the price of medical care in the USA push people to extensively rely on AI (or online communities btw), not that Americans are especially gullible or prone to AI psychosis.
FlashMobOfOne@lemmy.world · 2 pts · 7d
That's fair. Good point.
A_norny_mousse@piefed.zip · 8 pts · 10d
I clicked on the article to make sure: he used regular ChatGPT, not some specifically medical AI chatbot.
It's like with Fox "News": it's not news, it's entertainment. Taking it seriously is kinda dumb*. But at the same time, that's no reason to just let it fester and destroy people.
It's actually worse than Fox News. The companies are pushing all this not only like it actually works, but like it's the second coming of Christ. Which is criminally insane.
* or desperate
fullsquare@awful.systems · 2 pts · 10d
medical chatbot would be malpractice lawsuit pinata
A_norny_mousse@piefed.zip · 1 pts · 10d
I guess so; otoh I'm pretty sure some sort of AI-assisted healthcare exists.
FlashMobOfOne@lemmy.world · 7 pts · 9d
I'm of two minds about this.
Part of me wanted to blithely reply that "nothing of value was lost"; However, it's undeniable that this is a societal problem and this person is a victim every bit as much as someone who gets injured in a Tesla because it decides to run a red light and slam into a tree after a bad software update.
These products are built to be manipulative and harmful in the long-term, and by rights, should be regulated into oblivion. The LLM companies are just in the process of calibrating and tweaking the product so it kills more like cigarettes and less like interacting with a Terminator, and for that to happen, a few people have to die. They've almost certainly calculated out the potential litigation costs and accepted the risk.
And most of us can't avoid these products entirely because our job requires us to develop a basic proficiency in their use.
FE80@lemmy.world · 5 pts · 10d
AI psychosis is going to be great for culling the herd of morons.
Kirp123@lemmy.world · 4 pts · 10d
I'm surprised they don't have giant disclaimers to not believe medical advice from these bots. That seems much cheaper than paying lawsuits.
Zarobi@aussie.zone · 7 pts · 10d
They do, but you can't say "this isn't medical advice" and then turn around and give medical advice. Especially to the point of discouraging seeing a doctor.
magnue@lemmy.world · 2 pts · 10d
Man brings himself to the brink of death.
fullsquare@awful.systems · -1 pts · 10d
Man brings man to the brink of death, for profit