So only they can give direct legal and medical advice, which is still illegal in most countries. It's not a promoted service yet, but everyone is using it to ask such questions.
ChatGPT terms disallow its use in providing legal and medical advice to others
341–350 of 439 posts
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#342I read this not as "you can't ask ChatGPT about {medical or legal topic}" but rather "you can't build something on ChatGPT that provides {medical or legal} advice or support to someone else." For example, Epic couldn't embed ChatGPT into their application to have it read your forms for you. You can still ask it - but Epic can't build it. That said, I haven't found the specific terms and conditions that are mentioned…
(For anyone misunderstanding the reference to Epic, it's the name of an electronic healthcare record system widely used in American hospitals, not the game company Epic Games)
AI tends to be overly-positive, so either AI might congratulate you on your kill streak.
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#343Earlier quoted context omitted.
(For anyone misunderstanding the reference to Epic, it's the name of an electronic healthcare record system widely used in American hospitals, not the game company Epic Games)
AI tends to be overly-positive, so either AI might congratulate you on your kill streak.
> AI tends to be overly-positive, so either AI might congratulate you on your kill streak.
Nice
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#344Earlier quoted context omitted.
And this seems actually wildly reasonable. It’s actually pretty scammy to take people’s money (or whatever your business model is) for legal or medical advice just to give them whatever AI shits out. If I wanted ChatGPT’s opinion (as I actually often do) I’d just ask ChatGPT for free or nearly-free. Anyone repackaging ChatGPT’s output as fit for a specialized purpose is scamming their customers.
What if Chat GPT is used as a "power tool" for actual legal / medical professionals? E.G. Here's the set of documents uploaded by the customer. Here are my most important highlights (notice that document #68 is an official letter from last year informing the person that form 68d doesn't need to be filled in their specific situation). Here are my takeaways. Here's my recommendation. Apporve or change?
From the article
> The AI research company updated its usage policies on Oct. 29 to clarify that users of ChatGPT can’t use the service for “tailored advice that requires a license, such as legal or medical advice, without appropriate involvement by a licensed professional.”
I think your use case would probably be fine (although I am not a licensed professional, hahaha). I think they mostly want a licensed professional in the “driver seat,” or at least there to eat the responsibility.
It is sort of funny that coding agents became a thing, in this context. The responsibility-sink is irrelevant I guess because code writing doesn’t generate any responsibility as a byproduct.
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#345Earlier quoted context omitted.
And this seems actually wildly reasonable. It’s actually pretty scammy to take people’s money (or whatever your business model is) for legal or medical advice just to give them whatever AI shits out. If I wanted ChatGPT’s opinion (as I actually often do) I’d just ask ChatGPT for free or nearly-free. Anyone repackaging ChatGPT’s output as fit for a specialized purpose is scamming their customers.
What if Chat GPT is used as a "power tool" for actual legal / medical professionals? E.G. Here's the set of documents uploaded by the customer. Here are my most important highlights (notice that document #68 is an official letter from last year informing the person that form 68d doesn't need to be filled in their specific situation). Here are my takeaways. Here's my recommendation. Apporve or change?
In the legal context, AI is still prone to errors. And the legal system (in the US) doesn't have much in the way of guardrails.
Here's a case of several judges getting caught using AI to write error-filled judgements[1]. So it's not just lazy lawyers (several news articles over the past 3 years or so), but the courts themselves!
1 - https://www.reuters.com/sustainability/society-equity/two-fe...
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#346Earlier quoted context omitted.
"People with real questions get roped in by an enthusiastically confirming sycophantic model that will never question or push back, just confirm and lead on." That wasn't too buried IMHO
This isn’t accurate though, the models do push back and do question. I would say ChatGPT is way way better than the average therapist. I’ve seen maybe between 15-20 psychotherapists over the course of my life and ChatGPT is better than 85% of them I would wager. I’ve had a therapist recommend me take a motorbike trek across Europe (because that’s what he did) to cure my depression. I think people tend to radically ov…
>I’ve had a therapist recommend me take a motorbike trek across Europe (because that’s what he did) to cure my depression.
It's not bad advice.
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#347I have seen more than once people using ChatGPT as a psychologist and psychiatrist and self diagnosing mental illnesses. It would be hilarious if it was not so disasterous. People with real questions get roped in by an enthusiastically confirming sycophantic model that will never question or push back, just confirm and lead on. Tbh, and I usually do not like this way of thought, but these are lawsuits waiting to happ…
Completely agree. Let’s trust the experts we’ve reported on for years: put your symptoms into WebMD. Now please excuse me; according to WebMD apparently this stubbed toe means I have cancer so I have to get that treated.
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#348This feels like OpenAI tightening the language to catch up with how people are actually using ChatGPT, not a policy shift so much as a reality check
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#349Earlier quoted context omitted.
> LLMs hallucinate because training on source material is a lossy process and bigger, LLMs hallucinate because they are probabilistic by nature not because the source material is lossy or too big. They are literally designed to create some level of "randomness" https://thinkingmachines.ai/blog/defeating-nondeterminism-in...
So if you set temperature=0 and run the LLM serially (making it deterministic) it would stop hallucinating? I don't think so. I would guess that the nondeterminism issues mentioned in the article are not at all a primary cause of hallucinations.
Did you read the whole article?
"In other words, the primary reason nearly all LLM inference endpoints are nondeterministic is that the load (and thus batch-size) nondeterministically varies! This nondeterminism is not unique to GPUs — LLM inference endpoints served from CPUs or TPUs will also have this source of nondeterminism."
Re: ChatGPT terms disallow its use in providing legal and medical advice to others
#350[flagged]