[HN Gopher] Clinical knowledge in LLMs does not translate to hum...
___________________________________________________________________
Clinical knowledge in LLMs does not translate to human interactions
See also https://venturebeat.com/ai/just-add-humans-oxford-medical-
st...
Author : insistent
Score : 19 points
Date : 2025-06-14 22:18 UTC (41 minutes ago)
(HTM) web link (arxiv.org)
(TXT) w3m dump (arxiv.org)
| ekianjo wrote:
| > perform no better than the control group
|
| This is still impressive. Does it mean it can replace humans in
| the loop with no loss?
| majormajor wrote:
| What human? The control group was "instructed to instead use
| any methods they would typically employ at home." Most people
| don't have human-doctors-in-the-loop at home.
| jdiff wrote:
| No, the control group was instructed to "use any methods they
| would typically employ at home." So ChatGPT is no better than
| WebMD.
| ekianjo wrote:
| It's better as in, it's faster to give you an answer versus
| reading pages of WebMD
| brianpan wrote:
| You're wrong most of the time, but at least you get there
| quickly.
| dosinga wrote:
| Really what it seems to say is that LLMs are pretty good at
| identifying underlying causes and recommending medical actions
| but if you let humans use LLMs to self diagnose the whole thing
| falls apart, if I read this correctly
| majormajor wrote:
| Yeah it sounds like "LLMs are bad at interacting with lay
| humans compared to being prompted by experts or being given
| well-formed questions like from licensing exams."
|
| Feels to me like how two years ago "prompt engineering" got a
| bunch of hype in tech companies, and now is nonexistent because
| the models began being trained and prompted specifically to
| mimic "reasoning" for the sorts of questions tech company users
| had. Seems like that has not translated to reasoning their way
| through the sort of health conversations a non-medical-
| professional would initiate.
___________________________________________________________________
(page generated 2025-06-14 23:00 UTC)