https://www.npr.org/2026/04/30/nx-s1-5804474/ai-doctors-openai-patient-care-diagnosis
I’d much rather have doctors and other professionals (PAs, NPs) using AI tools to supplement their expertise, rather than pitting them against each other with the intent to eradicate one.
There are advantages to each. I think it’s best that each keeps the other in check.
I’m working through a foot issue and the doctor was going to consult another doctor and asked if he could consult AI.
I said sure. I’m waiting for it all to come back but he said he like to use the AI to check himself.
Who is training AI?
One of the teams my kid is on at his job uses AI to predict outcomes in healthcare, they get their information from actual data from patients. Most AI gets its data and information from the internet. IMO AI can help with healthcare, but will never replace doctors and other healthcare professionals.
The same people who share opinions of reddit and quora. Microsoft had an issue where their AI Twitter feed (that feeds other AI) began spewing antisemitic hate.
I am reminded of my good friend who was diagnosed with lung cancer several years ago. His daughter did a shitload of internet searches for lung cancer. As the doctor talked to the family, Lily took out a sheaf of printouts. The doctor looked at then and asked her to step out in the hallway, since none of those prognoses looked very promising. Jim said the doctor told her that everything she had was 100% correct except for one thing - this wasn’t the kind of cancer Jim had. He had renal cancer that had mestastized to his lung, and THAT is what they were going to treat. While he did have to have a lung removed, the remainder of the treatments was for renal cancer. That was 20 years ago. Jim finally did pass away late last year, but she got 20 more years with him.
AI is really a deep web search that the programs learn from. But when it comes to something like cancer, I’d rather have a doctor trained in the area have the final say on things.
I find it interesting he would ask. I’d think he’d use AI to give him ideas and then double check the details that the program found and run it by the established diagnostic methodologies.
Guessing that AI has not been addressed in the labyrinth of regulations and laws governing medicine, and getting the patient’s permission alleviates a liability concern.
I think technology is a great thing, as long as its users understand its limitations and take reasonable precautions to that effect.
Agreed. Reading the anecdote you shared, I think the system functioned exactly like it was supposed to.
Your friend’s daughter did as much homework as she could, and the highly trained and experienced professional revectored her well intended high energy and attentiveness with key information she was missing.
To your point, this story illustrates well why people should not go full rogue cowboy with serious, life-threatening illnesses. Homework is good, but informed medical opinions are still indispensable.
And if the medical opinion doesn’t smell right for whatever reason, people can always seek out the proverbial second opion.
I agree that AI is not to be relied on as the primary diagnostic tool although the physicians may use it for research.
as pointed out AI is trained by humans and AI can make mistakes.
I think for radiology it’ll became the standard very soon.
Indeed it is only as good as its programming protocols and the information made available to it.
I’m sure it gets most things right, but I feel there will always be a need for a human broker to sanity check the conclusions/recommendations.
Or even if eventually not truly a need, then perhaps a highly advised word to the wise.
A doctor needs to be in the loop for accountability.
For radiology it’s doing some impressive work. It’s finding things earlier with higher accuracy.
Recently I had a 12 year old brought to me with a history of pain for over 2 years- sometimes so unbearable she could not sleep. Numerous other dentists and physicians could not figure it out.
There was no clear-cut evidence on x-ray or anywhere else. I told the mother I had a good hunch, but nothing clear-cut to back my tentative diagnosis up.
I started a root canal on the tooth I suspected. It never had a filling, nor trauma, and there was no reason for the nerve to have become inflamed. But I took the gamble, and informed Mom that it was a calculated risk.
Next morning, the mother called to say the child was free of pain. I wonder if AI would have advised taking this risk with no evidence?
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.