4 ms·
I've used ChatGPT to help understand medical records too. It's definitely faster than searching everything on my own, but whether the information is reliable st
by Huxley1 11mo ago
I've used ChatGPT to help understand medical records too. It's definitely faster than searching everything on my own, but whether the information is reliable still depends on personal judgment or asking a real doctor.
More people are treating it like a doctor or lawyer now, and the more it's used that way, the higher the chance something goes wrong. OpenAI is clearly drawing a line here. You're free to ask questions, but it shouldn't be treated as professional advice, especially when making decisions for others.
- bushbaba 11mo agoI’ve found it just as accurate and a better experience than using telehealth or generalist doctors
- gxs 11mo agoDitto for lawyers The thing that gets me about AI is that people act like most doctors or most lawyers are not … shitty and your odds of running into a below average one are almost 50/50 Doctors these days are more like physicists when most of the time you need a mechanic or engineer. I’ve had plenty of encounters wher I had to insist on an MRI or on specific bloodwork to hone in on the root cause of an ailment where the doctor just chalked it up to diet and exercise Anything can be misused, including google, but the answer isn’t to take it away from people Legal/financial advice is so out of reach for most people, the harsh truth is that ChatGPT is better than nothing and anyone who would follow what it says blindly is bound to fuck up those decisions up in some way anyway On the other hand, if you can leverage it same as any other tool it’s a legitimate force multiplier The cynic in me thinks this is just being done in the interest of those professions, but that starts to feel a bit tin foil-y
- sarchertech 11mo agoYou can’t possibly have enough data to support that statement.
- jjtheblunt 11mo agothere was no mention of sample size, though, so the statement might be true for the commenter but not widely applicable, to your point
- sarchertech 11mo agoI’m saying that the commenter doesn’t have a large enough sample size to draw conclusions even for themselves. Just based on the amount of time ChatGPT has been out and the maximum number of different diagnosable and confirmable ailments they could have had in that time.
- terminalshort 11mo agoDo you have enough data to support the opposite conclusion? If not, your call for data is just a pathetic attempt to sound rational while making an empty appeal to authority.
- sarchertech 11mo agoI’m not calling for data. I’m saying the commenter doesn’t have a large enough sample size to draw conclusions—even informal ones. Just based on the amount of time ChatGPT has been out and the maximum number of different diagnosable and confirmable ailments they could have had in that time.
- ryandrake 11mo agoIf you're not a doctor, how do you know it's accurate? This is the huge problem with using LLMs for this kind of thing. How do you verify that it is better? What is the ground truth you are testing it against? If you wanted to verify that ChatGPT could do math, you'd ask it 100 math problems and then (importantly) verify its answers with a calculator. How do you verify that ChatGPT can interpret medical information without ground truth to compare it to? People are just saying, "oh it works" based on gut vibes and not based on actually testing the results.
- lp0_on_fire 11mo agoNobody would use these services for anything important if they _actually understood_ these glorified markov chains are just as likely to confidently assert something false and lie about it when pressed as they are to produce accurate information. These AI companies have sold a bill of goods but the right people are making money off it so they’ll never be held responsible in a scenarios the one you described.
- exitb 11mo agoIsn't statistical analysis a legitimate tool for helping diagnosis since forever? It's not exactly surprising that a pattern matcher does reasonably well at matching symptoms to diseases.
- matwood 11mo agoWhat is the most likely cause of this set of facts is how diagnostics works. LLMs are tailor made for this type of use case.
- sarchertech 11mo agoThat’s not what they’re trained on though and not what they’re doing. When you ask a general LLM like ChatGPT “I have a headache what is the cause” it’s going to give you the most statistically likely response in it’s training data. That might coincide with the most likely cause, but it’s also very possible it doesn’t. Especially for problems with widespread misconceptions.
- StarterPro 11mo agoYou should not be feeding your medical records into ChatGPT.
- matwood 11mo agoAnd you shouldn't be using Gmail or Google Search. At some point the benefits outweigh the costs.
- StarterPro 11mo agoHow does an email service compare to a glorified if-then machine that hordes all your data?
- eru 11mo agoWhy not? I can imagine a few different reasons you might have, but I don't want to guess.