ChatGPT and medicine’s future

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

ProRealDoc

Full Member
15+ Year Member
Joined
Jan 2, 2009
Messages
1,404
Reaction score
300

Can this technology replace some clinicians? Is it time to invest in AI?

Members don't see this ad.
 
Will it be as revolutionary as crypto has been???11

Seriously though just go over to it and ask it to answer a medical question. Pitch a case and ask it questions like you'd ask a resident. You'll quickly find you actually have nothing to be concerned about and that the people writing these articles know even less about medicine than hospital admins.
 
  • Like
Reactions: 7 users
Will it be as revolutionary as crypto has been???11

Seriously though just go over to it and ask it to answer a medical question. Pitch a case and ask it questions like you'd ask a resident. You'll quickly find you actually have nothing to be concerned about and that the people writing these articles know even less about medicine than hospital admins.
Yea but the admins I’ve experienced are only concerned about cutting cost and less about what is good for society.
 
  • Like
Reactions: 1 user
Members don't see this ad :)
I don’t see AI’s predictive clinical “skills” to be practical anytime soon. The problem with AI is that it doesn’t have clinical gestalt to know which variables or parameters are irrelevant. Each data point needs a weight associated with it that has to be programmed for it to compile to get to an answer.

I went to a talk a few years ago about AI in the OR and they ended with a real example where the team tried to make alarms practical and useful for hypotension and MH. The BP alarmed too frequently and was not clinical predicative. The MH alarmed whenever the EtCO2 made an increasing trend. Insufflation? Yup, it’s MH. Temp made an increase that it deemed was “too quick” even if it stopped at physiologic temp? Yup, MH. At least they were honest about their alarm goals and how it didn’t work yet

Granted this lecture was a few years ago and computer advancements move fast, but it was before that talk that Wilson won Jeopardy and people talked about you’d see Dr. Wilson in clinic. Hasn’t happened yet.
 
  • Like
Reactions: 1 user
I’m not concerned about it replacing anesthesiologists anytime soon but could see applications in radiology, pathology and primary care specialty areas where hands-on skills are not the norm.
 
Last edited:
We'll see AI eliminate diagnostic radiology as a specialty loooong before we see AI replace FP as a specialty.

The former, maybe in our lifetimes? (I plan to live a long time.) The latter? I doubt it.
 
  • Like
Reactions: 3 users

I find its utility to simply be its ability to parse, organize, and present pre-existing knowledge/information in a very natural way. Like a search engine on steroids.

Ultimately though, its language "ability" (i.e. write me a sonnet describing the mechanism of sugammadex in the style of Shakespeare) is nothing but a [very clever] parlor trick- not any kind of genuine "intelligence" as we understand the word.
 
Last edited:
  • Like
Reactions: 2 users
Top