
A new perspective published in JAMA challenges the growing narrative that artificial intelligence (AI) is poised to replace physicians, arguing instead that the technology exposes deeper structural failures in modern health care. “This is not really a story about the end of the doctor,” says Canio Martinelli, MD, OB/GYN, MSc, and lead author of the article. “It is a story about what kind of doctor, and what kind of medicine, we want to preserve and strengthen.”
What empathy scores really reveal
Recent findings show that AI can outperform physicians in perceived empathy during text-based evaluations by patients, fueling claims that even the most human aspects of medicine may be replicable. But Martinelli argues those results should be interpreted as a warning sign about what has happened within health care systems, not as evidence that machines have become compassionate.
The Viewpoint paper, “Artificial Intelligence Is Not the End of the Physician,” is written by researchers working with the Sbarro Health Research Organization (SHRO), under the leadership of Antonio Giordano, MD, Ph.D., Director of the SHRO, and Professor at Temple University.
“AI is sometimes described as outperforming humans even in areas such as empathic communication,” Martinelli says. “That should not be read as a triumph of machines, but as a signal of how far clinical practice has drifted from the conditions that make empathy possible.”
“Clinicians have been pushed further from the bedside and buried under expanding administrative demands,” adds Giordano, “and this has created a real fracture in the doctor-patient relationship.”
How AI could help restore care
Against this backdrop, AI’s apparent strengths reflect not a technological endpoint, but a systemic imbalance.
“AI should not be framed primarily as a tool to replace clinicians,” Martinelli said. “It should be framed as an opportunity to restore what medicine has progressively lost: time for reasoning, time for presence, time for human connection, and time for better teamwork around the patient.”
The article situates AI within the broader evolution of medicine itself, raising a pivotal question for health care systems: whether the technology will be used to further industrialize care, or to help free physicians from mechanical and bureaucratic burdens so they can become more thoughtful, more available, and more human in their work.
Clinicians’ role in shaping AI’s future
Martinelli also emphasizes that clinicians must play an active role in shaping how AI is integrated into care delivery.
“Clinicians should not remain passive observers of this transition,” says Martinelli. “Physicians need to be part of the conversation both as users of AI, and as people helping shape how it is implemented.”
Ultimately, the perspective calls for a reframing of the debate away from replacement and toward restoration.
“AI is not the end of the physician,” Martinelli says. “It is a test of whether medicine will use technology to replace human presence, or to finally make room for it again.”
“The real question is not whether AI can act like a doctor, but whether health care systems will use AI to absorb administrative tasks and restore the doctor’s uniquely human functions,” concludes Giordano.
Publication details
Canio Martinelli et al, Artificial Intelligence Is Not the End of the Physician, JAMA (2026). DOI: 10.1001/jama.2026.4356. jamanetwork.com/journals/jama/ … cle-abstract/2848376
Journal information:
Journal of the American Medical Association
Provided by
Sbarro Health Research Organization (SHRO)
The content is provided for information purposes only.
