NYC Hospital CEOs Push to Replace Radiologists With AI — and Say the Tech is Already Better Than Doctors
At a Crain's CEO forum, three of New York's top hospital executives called for regulatory changes to let AI read mammograms and X-rays without a radiologist present. The medical community is calling it dangerous.
Three of New York City’s most powerful hospital executives have done something unusual: they’ve gone on record saying they’d replace radiologists with AI today — if only the state would let them.
At a Crain’s New York Business CEO Forum panel on March 25, Dr. Mitchell Katz, CEO of NYC Health + Hospitals, told fellow hospital leaders he sees no technical barrier standing between current AI systems and performing radiology reads without a human in the room. “We could replace a great deal of radiologists with AI at this moment, if we are ready to do the regulatory challenge,” Katz said.
He then turned the question directly to his peers: “Is there any reason why we shouldn’t be pushing for changes to New York state regulations, allowing AI to read images without a radiologist?”
Two other CEOs answered with something close to a standing ovation.
Dr. David Lubarsky of Westchester Medical Center Health Network described his system’s experience deploying AI for breast cancer screening. “For women who aren’t considered high risk, if the test comes back negative, it’s wrong only about three times out of 10,000,” he said. “The AI Westchester uses misses very few breast cancers and is actually better than human beings.”
Dr. Sandra Scott, CEO of One Brooklyn Health — a smaller safety-net system running on tight margins — was more blunt. “I mean, I’m in charge of a safety-net institution,” she said. “It would be a game-changer.”
Radiologists Are Not Having It
The response from the medical community was swift and pointed. Dr. Mohammed Suhail, a San Diego-based radiologist with North Coast Imaging, called the CEOs’ comments “undeniable proof that confidently uninformed hospital administrators are a danger to patients — easily duped by AI companies that are nowhere near capable of providing patient care.”
“Any attempt to implement AI-only reads would immediately result in patient harm and death, and only someone with zero understanding of radiology would say something so naive,” Suhail told Radiology Business. He added a sharper critique: “Hospitals are happy to cut costs even if it means patient harm, as long as it’s legal.”
The tension is not new. Dario Amodei, CEO of Anthropic, made similar remarks about AI replacing radiologists in a podcast interview earlier this year, drawing a comparable wave of criticism from the medical field. The radiology community has consistently argued that AI systems, however accurate on average, lack the clinical judgment to handle edge cases — and that a missed diagnosis carries a human cost no error rate statistic fully captures.
The Regulatory Gap
What’s striking about Katz’s position is the framing: the technology, he implies, is ready. The obstacle is bureaucratic. He’s not wrong that AI has advanced rapidly. Of the more than 1,000 FDA-cleared AI-enabled medical devices on the market, roughly 75% are radiology applications. AI already assists with mammogram analysis, X-ray flagging, and stroke imaging at hospitals across the country. The question is whether it should operate autonomously.
New York state currently requires a licensed radiologist to review all diagnostic imaging. Changing that would mean amending state health regulations — a process that would inevitably involve medical boards, liability frameworks, and intense lobbying from both hospital systems and physician groups.
The regulatory environment is shifting in other ways. Gov. Kathy Hochul signed the RAISE Act in December 2024, establishing transparency and safety reporting requirements for AI developers and creating a state oversight framework to monitor risks tied to emerging technologies. Separately, Senate Bill S7263 would impose liability on companies whose AI systems impersonate licensed professionals — including healthcare — and provide advice that causes harm. Both laws suggest New York is moving toward tighter AI oversight, not looser rules for autonomous medical decisions.
The Workforce Problem No One Is Solving
Beneath the ideological battle lies a practical crisis. Radiologist shortages are well-documented, and demand for imaging services is growing faster than the workforce. At NYC Health + Hospitals — America’s largest public hospital system, serving over 1 million patients annually — the pressure is acute. Katz has led the system since 2018 and has a track record of pushing operational boundaries. His AI stance is consistent with that history: find cost-effective solutions, argue the regulatory case, let the politics follow.
Whether that approach survives contact with a radiologist community that feels genuinely threatened is another question. The CEOs made their bet in public. The regulatory response — and the medical community’s next move — will play out over the coming year.
