And medical diagnosis is one of the areas very well suited to automation, the only thing is answering some of the binary questions in the flowchart involve giving and interpreting complex tests. But a machine could decide what tests need to be done and would probably do a better job than a lot of doctors out there as long as they don’t add a bunch of conclusions like “patient is probably just looking for opiates” or “pain can’t be that bad if the patient is still able to function and talk about it” or “it’s never rare condition”.
Then the human doctors can be left to do the interesting work or find a new career if all they wanted to do is say whatever words will make the patient go away (and pay their bill).
Yeah, it wouldn’t need to be an llm or nn. But a flow chart questionnaire that can dynamically add new nodes can be considered a form of machine learning.
And medical diagnosis is one of the areas very well suited to automation, the only thing is answering some of the binary questions in the flowchart involve giving and interpreting complex tests. But a machine could decide what tests need to be done and would probably do a better job than a lot of doctors out there as long as they don’t add a bunch of conclusions like “patient is probably just looking for opiates” or “pain can’t be that bad if the patient is still able to function and talk about it” or “it’s never rare condition”.
Then the human doctors can be left to do the interesting work or find a new career if all they wanted to do is say whatever words will make the patient go away (and pay their bill).
People died because of their apathy. Deserved!
What you’re describing doesn’t even require AI. Skip that unreliability and just code a flow chart questionnaire.
Yeah, it wouldn’t need to be an llm or nn. But a flow chart questionnaire that can dynamically add new nodes can be considered a form of machine learning.