Every physician knows the feeling. You trained for years to think independently, to assess, to judge. Now you spend more than half your workday staring at a screen, entering data into fields someone else designed, responding to alerts someone else configured. The patient is in front of you. The system is between you.

That experience is not just a workflow frustration. It is a signal of something far more consequential that the profession has not yet reckoned with.

I have spent decades at the intersection of economics, AI strategy, and health care systems, advising health systems, payers, and clinical organizations on how technology reshapes institutional behavior. I have watched how industries convince themselves that transformative change is incremental until it is not. Medicine is making that mistake right now, and the consequences will be irreversible sooner than anyone in the profession appreciates.

The conversation about AI in health care is largely framed around tools. Better diagnostics. Faster documentation. Smarter decision support. Clinicians are processing AI the way they processed every previous technology wave, as something that assists, augments, and ultimately defers to human judgment. That framing is not just incomplete. It is dangerous.

AI is not a faster version of what came before. It is an altogether different category of thing. Every previous technology wave automated tasks. AI automates reasoning and, increasingly, decision-making itself. That is not an incremental advance. It is a categorical shift in what machines can displace. And it is not being understood or taken seriously enough.

But here is what is being missed entirely. The structural preconditions for physician displacement were not created by AI. They were created by electronic health records, and they have been accumulating for three decades.

EHRs restructured clinical thinking long before AI arrived. They did not just change documentation. They changed the architecture of how physicians think. Templates shape what questions get asked. Required fields determine what gets recorded. System-generated prompts direct attention before the physician has fully assessed the patient.

Research by Christine Sinsky and colleagues found that physicians spend more than half their workday on EHR tasks, with less than a quarter in direct patient contact. That is not a burnout problem. It is a cognitive restructuring problem.

Over thirty years, EHRs quietly transferred core cognitive functions from physician to system. The physician moved from primary evaluator to interpreter of a pre-structured record. That shift was already well underway before a single AI recommendation appeared in a clinical workflow.

AI-mediated EHRs are traditional EHRs on aggressive steroids.

AI systems are optimized for exactly the kind of environment EHRs have created. Structured. Standardized. Codified. As decision support tools grow more capable, every piece of clinical work formatted into a structured data field becomes a candidate for automation. The transition from human-led to machine-mediated decision-making does not require a dramatic leap. The groundwork has already been laid.

This creates what I call the validation trap. The physician’s role shifts gradually from decision-maker to validator of machine-generated recommendations. Not through a single dramatic policy change. Through the accumulation of individually unremarkable workflow decisions, each economically rational, each pushing in the same direction. Economic interests will accelerate this.

Health systems under margin pressure do not resist automation that reduces cost. Payers do not fund what cannot be measured. Technology vendors do not build for the preservation of physician autonomy.

There is also a compounding problem. Researchers at Oregon Health and Science University documented how errors embedded in the longitudinal EHR record persist and influence downstream clinical decisions in ways that are difficult to detect. In an AI-enabled system those errors are not merely carried forward. They are operationalized. The system encodes its own blind spots.

This is the uncomfortable truth the profession is not confronting. The window for intervention is open. It will not remain so indefinitely.

The path forward is not resistance to AI or retreat from EHRs. It is deliberate redesign of both around a fundamentally different objective. I have described this framework as Human-AI Synergy, grounded in my work on Recursive Human-AI Interaction. The question is not how much clinical work AI should perform. It is how physicians and AI can think together more effectively than either can think alone.

In Recursive Human-AI Interaction the loop does not close after a single exchange. The physician’s response to an AI recommendation becomes an input that refines the AI’s subsequent outputs. Each cycle produces a more refined clinical understanding than either party could have achieved independently. Properly designed, AI-mediated EHR systems can do the opposite of what current trajectories suggest. Not narrow clinical judgment. Vastly amplify it.

That outcome requires clinical AI systems designed to stimulate active evaluation rather than passive acceptance, with recommendations that are transparent, explainable, and explicitly challengeable. Physicians must remain the source of clinical intent, not reviewers of machine conclusions. Systems must amplify judgment precisely where human reasoning is most indispensable: ambiguity, ethical complexity, and individualized care.

The central challenge is not technical. It is institutional. Health care organizations, regulators, medical educators, and technology vendors are making deployment decisions today that will shape clinical practice for the next generation. Most of them are not asking the right questions.

This is not alarmism. It is pattern recognition from someone who has watched technology reshape industries that believed they were different.

Medicine is not different. It is just earlier in the process.

The question is not whether AI-mediated EHRs will continue to reshape how medicine is practiced. They will. The question is whether medicine will shape that transformation deliberately, or allow it to be shaped by economic incentives and system design defaults that nobody explicitly chose.

The window for drawing that boundary is open. It will not stay open indefinitely.

Matt Hasan is an economist, AI strategist, and founder of aiRESULTS. He advises health systems, payers, and life sciences organizations on the strategic implications of artificial intelligence, digital transformation, and emerging technologies. Over a career spanning more than four decades, he has held leadership and advisory roles with organizations including AT&T, IBM, Deloitte, Capgemini, and Citigroup, and previously served on the faculty of New York University’s Stern School of Business.

Dr. Hasan’s work focuses on the intersection of technology, institutions, and human decision making, with particular emphasis on how AI is reshaping medicine, governance, leadership, and professional practice. He is the founder of The AI Humanist Movement and an advocate for Human-AI Synergy, a framework that views AI not merely as a tool, but as a cognitive partner capable of extending human capabilities.

His writing includes “A Profession at the AI Frontier: Medicine Must Reinvent Itself or Cede Ground,” published in Health Affairs Forefront. He shares updates on LinkedIn and Medium.