Physicians need to demand a seat at the table in deciding how artificial intelligence is integrated into medicine. This is not a luxury or a branding slogan; it is a necessity. If physicians remain engaged, AI can become a powerful tool. If we disengage, we risk allowing others to define medicine on our behalf.
AI is already here. It is being used in documentation, diagnostics, risk prediction, imaging, scheduling, prior authorization, and claims review. The question is no longer whether AI will affect medicine. The real question is who will shape its use: physicians committed to patient care, or corporations focused primarily on scale, margins, and efficiency metrics.
We should be asking hard questions: Who owns the data? Who benefits from the model? Who is accountable when it is wrong? And are these tools being designed to support care, or to standardize it in ways that primarily serve administrators and payers?
AI is “coming for physicians’ jobs.” That framing is stale and simplistic. Medicine is not a checklist profession. It is a discipline built on ambiguity, nuance, ethics, and critical thinking. That is exactly what medical training teaches us. Medical school, residency, and clinical practice do not simply train us to memorize facts. They train us to evaluate uncertainty, recognize patterns while questioning them, weigh competing risks, communicate difficult truths, and make decisions in context. AI can process information at extraordinary speed, but it does not replace clinical reasoning. In fact, AI needs physicians’ critical thinking to be useful and safe. Moreover, AI does not replace the art of medicine.
Algorithms can suggest; they cannot truly understand. They can identify correlations; they cannot independently carry moral responsibility. They can summarize data; they cannot sit with a frightened patient, recognize what is unsaid, or appreciate the social realities shaping a treatment plan.
Increasingly, physicians worry that corporations are learning from our documentation, choices, and decision-making patterns to build systems that may later be used to monitor, direct, or even replace parts of our judgment. Is it too late to protect ourselves? No. But the window is open now, and it may not stay open forever. Physicians still have a chance to demand transparency in model training, governance over how clinical data are used, limits on surveillance-based productivity tools, and oversight of how AI recommendations are introduced into care delivery.
Being open to AI does not mean being naive about AI. It means approaching it honestly, critically, and transparently. When used responsibly, AI can help our patients. It can reduce clerical burden, streamline inbox management, improve access to information, assist with early detection, and free up more physician time for direct patient care.
Recent literature has reflected both promise and caution. Recent JAMA commentary has stressed that implementation must be clinically grounded, ethically governed, and focused on augmenting rather than replacing professional judgment. Some of the most troubling uses of AI in health care have not come from bedside clinicians, but from large insurers and corporations. A widely discussed 2023 lawsuit involving UnitedHealth alleged that an AI tool was used in ways that led to questionable Medicare Advantage coverage denials. When that happens, physicians are frequently left in the impossible position of explaining and absorbing the fallout with patients, making us look like the bad actors when we are not the ones building or deploying these systems.
This is why physician leadership matters now. This is a moment in technology to meet rather than be engulfed by. If we stay passive, AI will be layered onto medicine in ways that fragment workflows, erode autonomy, and widen the distance between doctors and patients. If we engage early, we can help build tools that respect how medicine is actually practiced. We still have time to control AI. We still have time to shape it into a tool that helps physicians care for patients more effectively, more thoughtfully, and more humanely. If we do this right, AI will not replace the art of medicine, it will remove obstacles that have pulled us away from it. And in doing so, it may help bring back what so many physicians and patients feel has been lost: the patient-physician relationship.
Sonal Patel is a pediatrician, neonatologist, and breastfeeding specialist whose work centers on infant nutrition, early development, and maternal well-being in the fourth trimester. In 2018, she founded NayaCare, a home health clinic dedicated to improving postpartum care.
She has written widely on maternal and infant health, with articles published in Scary Mommy, KevinMD, and The Colorado Sun. She is also the author of The Doctor & Her Black Bag, which explores maternal mortality through historical and personal perspectives while offering solutions to reduce it. A TEDx speaker, she presented The Economics of the 4th Trimester.
She cofounded and serves as co-executive director of the Center for 4th Trimester Care, a physician-led national nonprofit working to transform maternal health care. She also recently cofounded Pulse Med AI to bring physician voices into the AI digital space.
