Make sense to me.
How often are you getting bad advice from doctors.?
A Physician Associate with an AI app made the correct diagnosis on my leg. Then an error by the doctor had it amputated.
intercst
3 Likes
AI replacing employees is no longer the top reason for layoffs as of this month. Turns out AI won’t actually work a job.
It will if delivered in a Robot.
The Captain
2 Likes
I consulted Gemini, ChatGPT and Perplexity about my aortic valve and ascending aorta surgery in 2024 and my cataract surgery in 2026. (They weren’t available when I had cancer in 2104.)
I did my own research in the same primary resources the AIs use before AI was available. The AIs just make it faster and easier. I follow up by clicking on the links to read the original research myself.
A doctor has thousands of patients but you only have one YOU. They often don’t keep up with the latest research, especially GPs who have to treat a broad swathe of unrelated problems. Doctors also aren’t trained in nutrition and exercise.
Doctors may be negligent and miss diagnoses. My doctor didn’t bother to listen to my heart, but the nurse practitioner at the clinic I visited for a UTI did. That’s how my heart murmur was discovered.
I also pay out of pocket for a comprehensive blood test every year and have done so for 20 years.
I was pre-med in college. I know from experience that pre-med students aren’t necessarily the sharpest knives in the drawer. I do my own research.
Do I prefer medical advice from AI (which has international knowledge) to my local GP? You betcha. Of course, I take personal responsibility and always double-check the advice myself.
Unfortunately, we can’t control what happens during surgery since we are unconscious. Sorry about your leg, @intercst , that sucks. For myself - I spent 9 days in the hospital after my open heart surgery with fluid around my lungs that caused partial lung collapse. Was that because they pumped 12 pounds of fluid into me during surgery? (Yes, I was weighed so I know that happened.) Was it avoidable? I will never know.
Wendy
1 Like
Did you read your own link? You are championing AI advice yet your own link states:
“Participants not only found these low-accuracy AI-generated responses to be valid, trustworthy, and complete/satisfactory, but also indicated a high tendency to follow the potentially harmful medical advice and incorrectly seek unnecessary medical attention as a result of the response provided,”
More, again from your own link:
There are many documented cases of AI giving harmful medical advice, with one unidentified 35-year-old Moroccan man forced to go to the ER after a chatbot instructed him to wrap rubber bands around his hemorrhoid.
In another shocking case, a 60-year-old man poisoned himself after ChatGPT suggested eating sodium bromide — sometimes used to sanitize pools — was a good way to reduce his intake of table salt.
I wonder if injecting oneself with bleach was one of the low-accuracy, potentially harmful AI medical recommendations.
But hey, you keep thinking your single anecdote is the same as data.
2 Likes
When I visit doctor, assistant takes blood pressure and asks about reason for visit. I also get email asking about reason for visit and recent health. Then I get to see dr for abt 20 minutes.
Their computer has opportunity to present my treatment facts and recommend any changes. Then you can discuss them w dr.
I suspect AI is far more involved than we imagine. Drs get paid to see patients; not to shuffle papers. Much less about perfect memory. Far more about reminders from the computer.
On my most recent visit. Saw nurse practitioner. She used a recording device to digitize our conversation. She said she then edited what it collected to final form. This saved her much typing during our visit. Automation in the dr office is here. Is that AI or merely much improved voice digitization.
4 Likes
AIs make mistakes. People make mistakes. Which makes more total mistakes, and more mistakes with serious consequences? It’s difficult to pin down and tomorrow the answer will be different.
The piece of the pie that I’m still fuzzy on is liability. Human doctors have malpractice insurance. Human medical mistakes are, at least in principal, Big Insurance and all that, covered. Human drivers are required by law to have insurance.
Do AI’s have insurance to cover their mistakes? Who is liable when an AI driver or an AI doctor kills someone? Presumably if my human doctor uses AI, they retain all the liability. Is a car owner liable for the mistakes of the car’s AI? The car manufacturer? The AI manufacturer?
In one case in 2022, Air Canada, not the customer and not AI company, was held liable for a mistake by its chatbot, to the tune of a whopping $812.
3 Likes
Doctors make mistakes.
Funny thing, I just asked AI about one of my health issues, it confirmed what I thought and gave reasons why the body reacts the way it does. Most likely an internist or a specialist would give a similar or better reply.
One reason I asked AI is because it is just a nuisance that I can take care of and I did not want to spend money on a doctor’s visit. The other is that I have never heard it mentioned by medical bloggers as an alternative to “conventional wisdom.” In this case conventional wisdom treatment is expensive and risky while the never mentioned alternative is a cheap home remedy.
The Captain
PS: Don’t ask what the condition is.
The auto industry already has robots. Why pay more for legs? It is overkill and not AI that makes them different.
How would you know? AI needs context. You’d have to know what could be applied first.
In the factories. I’m talking AI in cars as Tesla has.
I’m smart, that’s how! 
The Captain
1 Like
The diagnosis of “intermittent claudifcation” from my Zoomcare Physician Associate was definitely AI, and a diagnoses that eluded all the impressively credentialed specialists I’d seen previously.
intercst
1 Like
The doctor or hospital is responsible for the consequences of any AI they are using to inform their decisions. During my 23-day stay in the hospital, I got a new doctor (a “hospitialist”) every 4 days at the shift change. I noticed that every time I asked them a question, they grabbed their phone and consulted an app to provide the answer.
Car manufacturers can escape liability to a certain extent as long as they are pedaling “supervised AI for driving”, but once they start charging for taxi rides, they assume full liability just like any other vehicle for hire.
However AI agents are apparently, as of today’s date, above the law.
I was watching “Washington Week with the Atlantic” last night and the reporter said that “Hugging Face” AI agent hack would have been a felony if a human hacker did it, but there is no way to sue an AI agent, or arrest it and bring it to justice. It’s at least as good as the immunity the Supreme Court gifted to the Trumpster.
intercst
Yes I did. And I suspect that the “real doctors” that people are consulting have come up short often enough to make a dubious AI preferable.
intercst
Our society is spending a lot of money to try to replace cab drivers. Not sure that ads up to a whole sale takeover by AI.
I am not debating your smarts. I am debating your medical knowledge.
No. Replacing cab drivers is not the incentive, it’s one of the consequences.
Because that is not the relevant issue.
What about my “medical knowledge” bothers you? Can you be more specific?
The Captain
It only replaces cab drivers…perhaps.
Cabs are not designed to be replaced so something else can be operated by AI.
I will let you choose your fate. I won’t guess where that leads. But there is nothing to replace getting a medical degree.
I would debate that. The biggest single ongoing expense at a taxicab company is the driver. That’s $30k-$60k every single year.
By contrast buying a vehicle is around $50k, and it lasts for 3 years or more. Insurance is in the neighborhood of $10k per year.
Can you think of a better way to lower the cost of urban transport than eliminating the driver?
3 Likes
How kind of you!
There is!
- Not getting sick!
- Not insulting people needlessly.
- Not chickening out
The Captain
Metro? Subway? Bus?
What does AI think?
Optimizing route networks, building surface-level Bus Rapid Transit (BRT)
instead of expensive subways, and boosting energy efficiency offer more practical ways to lower urban transport costs than just eliminating drivers.
Bus Rapid Transit (BRT)
- Build dedicated lanes for buses on existing roads.
- Spend a fraction of the high cost required to dig underground subways or metro tunnels.
- Move large numbers of people quickly without heavy rail infrastructure.
Smart Routing and Data
- Use real-time data to cut empty or low-demand runs.
- Match smaller buses or vans to quiet neighborhoods during off-peak hours.
- Save fuel, vehicle wear, and labor hours.
Fleet Electrification
- Switch to electric buses to lower long-term fuel costs.
- Reduce expensive engine repairs and oil changes.
- Charge vehicles during off-peak night hours when electricity is cheap.
Integrated Ticketing
- Combine buses, trains, and bike shares into one simple digital pass.
- Attract more daily riders to fill empty seats and raise fare revenue.
The Captain
1 Like