‘Unbelievably dangerous’: experts sound alarm after ChatGPT Health fails to recognise medical emergencies
http://www.theguardian.com/technology/2026/feb/26/chatgpt-health-fails-recognise-medical-emergenciesOpen linkView original on feddit.uk182
Comments12
I'm sorry. But if you are following chatgpts medical advice... It's the kind of thing that will take care of itself.
It’s not like a user asked vanilla ChatGPT random questions and took the responses for medical advice. This is a service openAI marketed specifically for medical use. If we want to keep up the pretense that consumer protection laws and regulations exist and matter, this is a big deal, and the blame is on the vendor, not on the user.
I'm guessing they didn't have any other choice.
Yeah, a lot of people are short on money to swing a doctor visit and whatever costs that may hold.
They'll get sued and will have to prevent it from dispensing medic advice. Honestly I'm surprised they weren't sued already.
Or they get a pass like so many corporations are now. Like self driving cars.
I don't recall Cruise getting a "pass."
This is actually criminal activity, it poses as qualified expert advice, but is nothing but a con, delivering totally unqualified advice.
This can only be categorized as fraudulent, in any lawful society where quackery is illegal.
ChatGPT needs to be investigated, and legal procedures needs to be held against them, with the goal of closing the "service" and huge fines for putting lives at risk.
am i in an anti-ai bubble or is every ai story that comes out extremely negative?
Definitively a bubble. I know pro-AI people irl
Personally I'm more in the "it's not the tool, but how it's used (aka, the rich fucks are the problem, as usual)"
Bubble, mostly. I'm sure you can find plenty of pro-ai articles elsewhere.
New flash! AI chatbot is not a doctor! More news at 7.