Hospital executives see artificial intelligence as a cost-saving game-changer for medical imaging, but radiologists warn that sidelining human experts could put patient lives at risk.
- Executive Push for Automation: Top administrators at major New York hospital systems are openly advocating for regulatory changes that would allow artificial intelligence to perform initial reads on X-rays and mammograms without a human radiologist.
- The Promise of Efficiency: Proponents argue that AI is already highly accurate—missing very few breast cancers—and that delegating first reads to machines could produce massive financial savings and expand screening access, particularly for cash-strapped safety-net hospitals.
- Fierce Clinical Backlash: Practicing radiologists are sounding the alarm, accusing hospital administrators of being dangerously uninformed. They argue that implementing AI-only reads would prioritize legal cost-cutting over patient safety, inevitably resulting in medical harm.
The intersection of healthcare and technology has long been a battleground between administrative efficiency and clinical caution. Today, that battle line is being drawn directly over the radiology department. As artificial intelligence continues to evolve at breakneck speed, hospital administrators are beginning to view the technology not just as an assistive tool, but as a viable replacement for human specialists in certain clinical circumstances.
Mitchell H. Katz, MD, president and CEO of NYC Health + Hospitals—America’s largest public hospital system, which he has led since 2018—made a striking declaration. Acknowledging the rising demand for medical imaging and the surging costs associated with the human experts who interpret them, Katz stated that he is prepared to let AI take the reins. “We could replace a great deal of radiologists with AI at this moment, if we are ready to do the regulatory challenge,” Katz told the forum. His vision involves a workflow where artificial intelligence handles the “first reads” of mammograms and X-rays to increase access to breast cancer screenings. Human radiologists would only step in to double-check scans that the AI flags as abnormal, a shift he believes could produce major savings.
Katz is not alone in his optimism. Fellow panelist David Lubarsky, MD, MBA, president and CEO of the Westchester Medical Center Health Network, reported that his system is already seeing immense success with AI deployment. According to Lubarsky, the AI utilized by Westchester is “actually better than human beings” at avoiding missed breast cancer diagnoses. He noted that for women who are not considered high-risk, an AI’s negative test result is incorrect “only about 3 times out of 10,000.” Bolstered by these statistics, Katz openly questioned whether hospital leaders should begin lobbying to change New York state regulations to permit AI to read images entirely without a radiologist. For smaller institutions facing tight financial margins, this administrative shift represents a lifeline. Sandra Scott, MD, CEO of One Brooklyn Health, agreed with Katz’s push. “I mean, I’m in charge of a safety-net institution,” she noted. “It would be a game-changer.”
The enthusiasm in the boardroom stands in stark contrast to the outrage in the reading room. Radiologists argue that the technology, while impressive, is nowhere near capable of operating independently of human oversight. The tension mirrors recent industry fallout involving Dario Amodei, PhD, CEO of the AI company Anthropic. In a recent podcast interview, Amodei falsely stated that artificial intelligence had already taken over the core functions of radiology, sparking widespread condemnation from medical professionals who felt their complex, nuanced work was being fundamentally misunderstood by tech executives.
When clinical experts heard Katz’s recent comments, the pushback was swift and severe. Mohammed Suhail, MD, a San Diego-based radiologist with North Coast Imaging, offered a blistering critique of the administrative eagerness to automate his specialty. “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,” Suhail told Radiology Business. He warned that any attempt to implement AI-only reads would immediately result in patient harm and death, characterizing the executives’ perspective as hopelessly naive. “But in some sense, they’re correct,” Suhail added. “Hospitals are happy to cut costs even if it means patient harm, as long as it’s legal.”
As healthcare systems grapple with post-pandemic financial strain and soaring labor costs, the allure of an AI workforce will only grow stronger. Yet, as the standoff between hospital CEOs and clinical radiologists demonstrates, revolutionizing modern medicine requires more than just highly capable algorithms. It requires navigating a profound divide between those who balance the spreadsheets and those who hold patient lives in their hands.