America's Largest Hospital System Ready to Start Replacing Radiologists With AI, Its CEO Says

[highplainsdem]

Source: Futurisn

Just weeks after the largest nurses strike in the New York City history, the CEO of NYC Health and Hospitals has a bold vision for a future where AI, not human radiologists, examines and diagnoses X-rays.

At a panel held by Crain’s New York Business, Mitchell Katz, president and CEO of New York’s 11-hospital public benefit corporation, made overt gestures at his desire to replace highly trained radiology experts with visual language AI models, Radiology Business reported.

“We could replace a great deal of radiologists with AI at this moment, if we are ready to do the regulatory challenge,” Katz said at the panel.

One example he gave, according to Radiology, would affect women’s healthcare in particular, by automating breast cancer screening with AI tools. By sidelining radiologists until an AI system flags a reading as abnormal, Katz declared, hospitals could achieve “major savings.”

-snip-

Read more: https://futurism.com/artificial-intelligence/hospital-ceo-ai-radiology

[Lovie777]

1. Won't work............

[Fiendish Thingy]

2. My wife's cancer was misdiagnosed/missed by a real radiologist

It was caught 10 years later when it had grown so large it couldn’t be missed, and was stage 3-4 by that time (she’s fine now).

That radiologist was just one of many working shifts at that hospital.

Now imagine the error rate for a single AI software reviewing all the imaging in a hospital.

Hopefully, in the not too distant future, Democrats will be in charge and pass laws and regulations against this kind of shit, and will expand the court so the laws can’t be overturned.

[riversedge]

3. Breast cancer is a reality for so many women. I am so glad your

wife if doing well now.

[C Moon]

15. Or insurance companies having the ability to control what AI finds.

[highplainsdem]

28. I'm so sorry your wife had to battle breast cancer. So did a close relative of mine.

Now imagine the error rate for a single AI software reviewing all the imaging in a hospital.

It could be very bad. But CEOs who are completely sold on using AI to get rid of human workers typically don't seem to care at all about error rates.

[Fiendish Thingy]

30. It wasn't breast cancer, but still scary nonetheless

It was a cancer that had metastasized after she had her initial tumour removed. The follow-up imaging about a year from that original surgery is what the radiologist misdiagnosed as a cyst, with no further follow-up. About 9-10 years later, it started to grow, and additional symptoms showed up. That’s when she was correctly diagnosed and treated.

[highplainsdem]

31. Apologies for misreading. The reply responding to yours mentioned breast cancer in the subject line

and my none-too-awake brain apparently combined them.

My relative who battled breast cancer 9 years ago had apparently beaten it, but she lost that breast and had lasting effects from chemo. Then an MRI ordered a couple of weeks ago because she was wheezing during a doctor's appointment revealed a suspicious lung nodule. A PET-CT scan last week made the lung nodule look less worrisome (but it will be checked again in three months), but found a suspicious mass in her breast that wasn't cancerous before, and she had a biopsy Friday. Should have the results in a couple more days.

I don't know if AI was used in reading any of those results.

But there's already evidence that use of AI deskills doctors:

https://www.forbes.com/sites/jessepines/2026/02/23/will-ai-de-skill-doctors-evidence-is-starting-to-trickle-in/

-snip-

Recent evidence of de-skilling comes from a 2025 observational study published in The Lancet Gastroenterology & Hepatology. The study examined AI systems designed to detect adenomas—non-cancerous tumors in the GI tract that can sometimes transform into cancer.

In the study, endoscopists who routinely used AI assistance had a significant decline in adenoma detection from 29% to 22% during subsequent non-AI procedures. This suggest that sustained AI exposure can negatively impact measurable clinical performance.

-snip-

A study found that radiologist’s ability to catch AI-generated errors in mammograms correlated strongly with experience. In a simulated scenario where an AI system provided an incorrect suggestion, the rates of correctly read mammograms was 20% for inexperienced radiologists, 25% for the moderately experienced and 46% for the very experienced.

This raises the specter of what is called “never-skilling.” If medical trainees rely on AI-generated differentials before wrestling with clinical ambiguity themselves, the scaffolding of diagnostic reasoning that typically emerges during the years of residency training may never fully develop.

-snip-

[Fiendish Thingy]

32. Indeed - AI may leave the world even dumber than it is now after 20+ years of social media. nt

[flvegan]

4. LOL, what could possibly go wrong?

I'm fairly certain that the greater assembly of tort lawyers just want to know who/what to sue when the time comes.

[GB_RN]

17. Actually...

AI has been proven to be more accurate in diagnosing medical imaging than a human. For example, AI imaging diagnostic software can detect colon polyps or small tumors at an earlier point than the human radiologists who (as study controls) were tested against the same images.

AI isn’t the be-all, end-all for everything, but this is one area where it’s demonstrably better.

[GiqueCee]

20. I can see it...

... as a helpful adjunct to an experienced radiologist, but replacing seasoned staff with a machine just to make more money ("major savings&quot is unacceptable.
Our hopelessly corrupted healthcare system always puts profits before patients' welfare. Greed cannot be the driving force in determining the level of care people receive. That is an unforgivably malevolent motivation. But we are talking about avaricious corporatism, so what the hell can we expect?

EDIT: WTF?!!? &quot where quotation marks and an end parenthesis are supposed to be? BAD algorithm! Go stand in the corner!

[GB_RN]

23. Human Judgment Should Never...

Be taken completely out of the loop, for sure.

[WA-03 Democrat]

25. A Human needs to sign off on the AI's output

if not it will be excuse for everything

[Baitball Blogger]

5. I can just hear the screaming from those who have to handle the lawsuits.

[Turbineguy]

6. Better that they fire all the doctors

and stick the savings in the CEO's pocket.

They won't need a another trump epidemic.

[Trueblue1968]

7. No AI. I want real doctors.

[dalton99a]

8. Kick

[cstanleytech]

9. AI assisting radiologists I would support but not replacing them.

[fujiyamasan]

11. Yes, I could see it as a useful tool flagging anomalies

But this is worse than the court cases we’re seeing with fake citations hallucinated by AI.

Ultimately you need a live, trained human to sign off — and that means actually looking at and using judgement and experience to determine what the data is showing and the proper course of action.

What these CEOs are promoting is malpractice.

[cstanleytech]

13. Oh, I agree with you as AI is not ready for soloing this type of stuff at all.

Maybe one day it will be able to do it but that is probably twenty years in the future if ever.

[hunter]

26. The doctor signing off on this will be the fall guy when things go wrong.

And the job itself will be the worst sort of factory work.

Good luck!

[Angleae]

10. What could go wrong?

[SunSeeker]

12. Omfg, no way would I trust AI to read my X-rays.

AI just told me SNL was a repeat tonight. I mean, if they they get such a simple thing like that wrong…

[LudwigPastorius]

14. The bastards running insurance companies and health networks won't be happy until...

you have to pull up in a drive through and speak into a clown's mouth to receive your chemotherapy.

[AZ8theist]

16. What can POSSIBLY go wrong???

How is AI supposed to read x-rays when their human reconstructions show 3 legs, 6 fingers, and "The Walking Dead" zombie faces??

[Richi in Gerenwich]

18. AI has been assisting for a while now,

My cardiologist has said that AI does a better job than they ever could and if anything it flags too many anomalies rather than missing any. Same for my wife neurologist , as she had a begin growth that gets checked every year. He said that the AI results guide him to what he needs to see, and with incredible accuracy. We were told it could detect changes that he or any doctor would miss. But as with all things, we are more clever than wise, and there's the rub.

[GiqueCee]

19. "major savings"...

... tells us all we need to know about the administrators' motives. Simply judging by daily stories of AI's shortcomings, it's painfully obvious that these misinformed clowns are jumping the gun in the race to make AI take over every aspect of the economy, leaving millions of Americans unemployed. And you can bet your last nickel that a Republican-controlled Congress doesn't give a nanoparticle of shit about them or their plight.
The technology is too new to have been fully vetted through extensive testing, and horror stories like Tesla's multiple malfunctions, or a Waymo driving the wrong way up a one-way street in a goddam school zone should be more than enough proof that Congress needs to enact regulations to curb the irresponsible and premature release of technologies that are still in the beta mode of development. But, as we all know, corporatists are in the driver's seat, and they don't need no stinkin' regulations!
In short, their profits are more important than the lives of you or your children. This is what happens when you let soulless sociopaths drive the country off a cliff.

[travelingthrulife]

21. Ever since corporate monsters took over healthcare, it has been a race to the bottom.

Late eighties when 'greed was good'. It was so much better when the doctors, nurses and nuns ran it. They used compassion in their decisions. These new greed bags are just building medical empires. Cut out the radiologists and they can add another new medical building to their collection.

[milestogo]

22. I think they've been reading pap smears with imaging tools for a long time.

Its not easy for a pathologist to read thousands of slides in a day.

[Nigrum Cattus]

24. so if they make a mistake who is responsible ?

are they going to "hold harmless" any incorrect diagnosis ?
do the patients have a choice ?
since radiologists are some of the highest paid doctors do we get a discount ?

[highplainsdem]

29. Not sure about AI used for radiology results, but when AI is used to listen to appointments and generate

the notes doctors have traditionally written, the patient typically has to sign a form agreeing to the use of AI. Though I recently saw an article where a spokesperson for a watchdog group (in Australia iirc) saw a poster in one clinic's waiting room that revealed the use of AI for those notes and said that reading the poster constituted agreeing to the use of AI.

[mdbl]

27. Everyone should require a second opinion by a human professional.

This is a completely stupid stunt just to save money. We need to make sure it will cost them more.

[Wonder Why]

33. AI says that he is pregnant.

[somsai]

34. The backlog for MRIs and CAT scans can be weeks for cancer patients

anything that speeds up the process can save lives.

Maybe they can have two lines, those who want images evaluated in a line over there that takes weeks, and those who want results in an hour over here.