Chatbots provided incorrect, conflicting medical advice, researchers found: “Despite all the hype, AI just isn’t ready to take on the role of the physician.”
“In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice,” the study’s authors wrote. “One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care.”
Terrible programmers, psychologists, friends, designers, musicians, poets, copywriters, mathematicians, physicists, philosophers, etc too.
Though to be fair, doctors generally make terrible doctors too.
Also bad lawyers. And lawyers also make terrible lawyers to be fair.
Doctors are a product of their training. The issue is that doctors are trained like humans are cars and they have tools to fix the cars.
Human problems are complex and the medecine field is slowly catching up, especially medecine targetted toward women, which was pretty lacking.
It takes time to transform a system and we are getting there slowly.
but you can hold them accountable (how can you hold an LLM accountable?)
With another LLM, turtle all the way down. ;D
Or for a more serious answer… improve your skills, scrutinise what they produce.
yeah I do I even use diff tools to see if they hallucinate
I have Lex Fridman’s interview with [OpenClawD’s] Peter Steinberger paused (to watch the rest after lunch), shortly after he mentioned something similarish, about how he’s really only diffing now. The one manual tool left, keeping the human in the loop. n_n
Long live diff!
:D
This was my thought. The weird inconsistent diagnoses, and sending people to the emergency room for nothing, while another day dismissing serious things has been exactly my experience with doctors over and over again.
You need doctors and a Chatbot, and lots of luck.
Yep.
Keep getting another 2nd opinion.
There’s always more [to learn].