A rare disease is not necessarily hard to diagnose if it has distinctive symptoms, as opposed to symptoms that could easily be one or more much more common diseases.
I am guessing that the way forward is for people to use this technology as a tool. Not as a replacement for anything, be it writing essays or diagnosing disease.
Isn’t this what has been envisioned for many decades? A world where the computers are spoken with and they give us answers to our problems. I’d say our technology is finally catching up to our dreams.
The computer has been fed all the data it needs to make a diagnosis, in some ways similar to how a doctor would have been fed information, crammed in. The computer then analyzes the symptoms and testing inputs to come up with a logical theory. It actually doesn’t seem like “rocket science” to me. (Another thing that computers are good at)
I work in electronics. Troubleshooting issues is a major skill that is extremely useful and can be gained through study and experience, not completely unlike medical diagnosis. Given the correct “symptoms” a computer could certainly reason through the same process.
Seems like the doctor might be more concerned that his expertise is being replaced in some way by a computer program, making some of his skills less “specialized”. Other mediocre doctors, or even patients, may be able to reach the same conclusion just as easily. Many already do, scouring the internet for answers. Obviously this could be threatening to his professional status on some level.
No worries doctor, you’re still necessary, for now. We’ll let you go ahead and certify the diagnosis so you can feel important. It’s an interesting time to be alive. Big change is on the way.
I’m no expert in medicine, but I suspect that a rare disease would be easy to be misdiagnosed. The Harvard doctor chose this particular rare disease presumably because it was hard to diagnose. The article only mentioned this one real-life case, but I’d assume the doctor had run many other cases through the AI system to reach his conclusion that “GPT-4 has better clinical judgment than ‘many doctors’”.
Yes. An incredible percentage of our GDP (nearly a fifth and growing) is spent on healthcare. A solution may be on the way, as we can’t seem to solve the problem any other way. This technology could dramatically reduce the cost of healthcare.
This seems more an indictment of doctors.
This is a book promotion then.
PS. I am sure that AI can help in medical diagnosis. But what data to feed it is also major consideration that the doctor elides over. And then we all know the capability of chatgpt to make things up. How do we trust what it comes up with?
I wouldn’t assume anything because as you know …
The idea that technology will be our savior is very attractive and very popular in the US. Meanwhile from healthcare to infrastructure to college everything is extremely expensive here compared to other parts of the world.
How do we trust what a doctor tells us? No one (or machine) is always right. If a machine is right more often than a typical doctor, that’d be a sea change.
Initially I see it as a doctors tool. A doctor will still ask the questions and use AI to diagnose and recommend further treatment. Assuming the doctor is in agreement the treatment will proceed. After a fair amount of A/B testing I can see the AI doing the diagnosis and a GP merely confirming it. We will need far fewer GPs then. Eventually I see the work of a GP being done by AI and it will move into more specialized forms of medicine. The doctors will be needed to confirm, communicate and be an integral part of all surgeries and procedures. The area I don’t see being affected is nursing. I think nurses will take on a more important roll as AI becomes more important. Just my crystal ball. I could be very wrong.
I don’t see this as being much different than individuals using the internet trying to self-diagnose, then going to their doctor more informed about the possibilities.
Anything that can help get to the correct problem is a plus in my book. Then hopefully there are workable treatments!
But maybe that’s just me.
Now imagine it was not a lawyer but a doctor.