Doctors are great at a lot of things but have never been good at accurate diagnosis when you look at the data
I am very open to whatever lets human doctors do what they do best & supports them with possible/likely diagnosis to consider
As a person who had a misdiagnosed rare cancer and caught the misdiagnosis myself by going to a research library and using their medical journals, I see the benefit of AI second opinions even if your doctor is pretty sure about whatever they diagnose
I would be comfortable with CaBOT being an adjunct, and even more comfortable if it was blinded. My concern would be it becoming a crutch.
I don't think we should allow general skepticism in AI (which I share btw) to blind us to specific use cases where RAG/LLM are doing things which can be shown to be as good as or better than an expert in role, at least the expert you can afford to see.
One downside of improved imaging has been an increase in "detected marker for..." causing patients to present. Not all of these need to. It takes time to learn how to nuance improved diagnostics and I think CaBOT will be part of that on both the good and the bad side.
Whoever you bought the AI diagnosis service from, and, under standard defective product liability rules, anyone in the chain of commerce back from there up to and including the manufacturer
Doctors are great at a lot of things but have never been good at accurate diagnosis when you look at the data
I am very open to whatever lets human doctors do what they do best & supports them with possible/likely diagnosis to consider
As a person who had a misdiagnosed rare cancer and caught the misdiagnosis myself by going to a research library and using their medical journals, I see the benefit of AI second opinions even if your doctor is pretty sure about whatever they diagnose
I would be comfortable with CaBOT being an adjunct, and even more comfortable if it was blinded. My concern would be it becoming a crutch.
I don't think we should allow general skepticism in AI (which I share btw) to blind us to specific use cases where RAG/LLM are doing things which can be shown to be as good as or better than an expert in role, at least the expert you can afford to see.
One downside of improved imaging has been an increase in "detected marker for..." causing patients to present. Not all of these need to. It takes time to learn how to nuance improved diagnostics and I think CaBOT will be part of that on both the good and the bad side.
"It's always Rhabdo" (mandatory House quote)
(not a doctor)
https://archive.ph/pGY5G
What are doctors for?
Who are you going to sue if the AI diagnosis is wrong?
Whoever you bought the AI diagnosis service from, and, under standard defective product liability rules, anyone in the chain of commerce back from there up to and including the manufacturer
Good luck with that.
Whoever sold you the AI diagnosis was a criminal --- providing medical services without a license.