How long until many doctors are replaced with super computers?

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
I routinely use the most advanced CAD(Computer Aided Diagnosis) software available when reading mammography. I have find that the performance of these systems to be quite pathetic. Based on what I have seen, I do not see CAD replacing radiologists in our life time. The task of the Radiologist is much fussier than most non radiologists think. Things that are very simple for humans like finding an appendix on a CT scan of the pelvis are impossible for the computer even using the most current advanced AI technologies. The appendix looks different in every patient and can look different in the same patient at different times. The appendix can be very difficult to distinguish from small bowl. This makes it very difficult to train a compute to find an appendix. This is a very simple task that the computer has great difficulty with. When reading CT scans, radiologists deal with much more difficult tasks, such as distinguishing fluid in bowel from an abnormal fluid collection such as abscess. There is infinite variability in the appearance of the small bowel on CT. This makes it very difficult for CAD to do find abnormal fluid collections in the peritoneal cavity.
 
Last edited:
Those devices are still operated entirely by physicians. The robotic apparatus is just a tool. It does no thinking and doesn't have any autonomous function.

I was going to rebut, but Naylor took the words out of my mouth.
 
This thread is full of fail from 19-year-olds who know how to "hack facebook" AKA predict computer AI trends in medicine in the next 100 years.

Here's what I think: computers are great at storing a LOT of information, and this ability is doubling at a crazy fast rate. Computer algorithms are only as great as the programming team behind it, so while a computer may not be the ultimate diagnostician, it will be able to take massive amounts of real-world data and decide on a statistically acceptable path of treatment. No doctor is able to read every new article in the scientific journals, let alone Up To Date, every day and run a practice. Computers will help with this. Surgery won't happen anytime soon, there's no way. And HAL 9000 won't put you under with ketamine and reduce your shoulder with the bedsheet. But think about radiological imaging: You feed an AI millions upon millions of CT's and MRI's, and it'll create a statistically probable map of what a human body should look like. Just like a seasoned radiologist can spot something wrong on a film, an algorithm can be trained to do the same thing. Anyone with photoshop knows that the first healing brush was a joke, and it was an art just using the thing. CS5's healing brush is nothing short of magic, making some amazing context-sensitive calls on what the "background" is supposed to look like. This is after only a decade of photoshop being out. Think about what 50 years in radiology advances will bring, especially with healthcare dollars basically up for grabs in the upstart tech market.
 
i could see computers coming to replace certain fields or largely replace some fields, but i don't see them taking over any time soon.

1) to a LOT of patients, having good interactions with a doctor is very very important. i highly doubt that even i you convince them that the computer is smarter/better, they will be leaping to replace a human doctor with a computer one.

2) someone else brought this up, but what if the machine messes up? who is liable for that legally? do we sue the creator of the machine? do we sue the hospital for recommending using the machine over a human? etc. i think it would take a lot of time to work this out too even if we are able to get infallible machines
 
See:
<---------------------
i could see computers coming to replace certain fields or largely replace some fields, but i don't see them taking over any time soon.

1) to a LOT of patients, having good interactions with a doctor is very very important. i highly doubt that even i you convince them that the computer is smarter/better, they will be leaping to replace a human doctor with a computer one.

2) someone else brought this up, but what if the machine messes up? who is liable for that legally? do we sue the creator of the machine? do we sue the hospital for recommending using the machine over a human? etc. i think it would take a lot of time to work this out too even if we are able to get infallible machines

Thank you for these new and unique contributions to the debate.
 
Lawyers, pharmacists, nurses, dentists, fire fighters, pilots, soldiers, and home makers would be replaced before doctors, so we will win since we are going to be laid off later than them.
 
I honestly hope not any time soon or otherwise my time is wasted trying to become a doctor in the first place.🙂
 
Top