By Matt Hasan, PHD

Artificial intelligence is forcing medicine to make a choice that has received surprisingly little attention. Should AI be designed to replace increasing portions of clinical judgment? Or should it be designed to elevate it? Much of today’s discussion assumes these are the same objective. They are not. One seeks to substitute human reasoning. The other seeks to strengthen it.

Unfortunately, the momentum today is largely toward replacement. We read almost daily about AI systems matching or even exceeding physician-level performance in interpreting images, generating differential diagnoses, recommending treatments, and communicating with patients. The implicit assumption is that as these systems become more capable, the physician’s cognitive role will steadily diminish.

I believe that is the wrong destination. The real opportunity is not to build AI that thinks instead of physicians. It is to create AI that helps physicians think better than either humans or AI can think independently. That requires a fundamentally different model of physician-AI collaboration.

For generations, medicine has treated clinical judgment as an individual accomplishment. We celebrate the physician who recognizes the subtle diagnosis, asks the decisive question, or notices what everyone else overlooked. Clinical excellence has been synonymous with the ability of a highly trained individual to synthesize information, weigh competing possibilities, and make sound decisions under uncertainty. AI allows us to rethink that model.

Most current applications treat AI as either a sophisticated search engine or a clinical consultant that produces recommendations for physicians to accept or reject. The interaction is typically linear. The physician asks a question. AI generates an answer. The physician decides what to do with it. That is useful, but it barely scratches the surface of what is possible.

I believe AI should participate in a recursive process of clinical reasoning in which physician and AI continually challenge, refine, and expand each other’s thinking. Rather than a single exchange, the interaction becomes an ongoing intellectual dialogue. The physician questions the AI’s assumptions, asks it to reconsider conclusions in light of contextual information, and explores alternative explanations. AI, in turn, identifies overlooked possibilities, surfaces relevant evidence, and challenges the physician’s initial impressions. Each iteration improves the next.

This concept forms the basis of my ongoing work on Human AI Cognitive Synergy (HACS). HACS proposes that recursive collaboration between physicians and AI has the potential to produce forms of diagnostic reasoning and clinical judgment that neither could achieve independently. The central premise is straightforward. The goal should not be better artificial intelligence. The goal should be better clinical intelligence.

Whether that hypothesis ultimately proves correct remains an empirical question. It deserves rigorous scientific investigation rather than assumption. But I believe it points medicine toward a far more promising future than one centered primarily on replacing physician cognition.

If AI is designed principally to automate diagnosis, physicians inevitably become supervisors of increasingly autonomous systems. Over time, clinical reasoning itself risks becoming progressively outsourced. History suggests that cognitive skills decline when they are no longer routinely exercised.

If, instead, AI is intentionally designed to deepen physician reasoning, the relationship changes fundamentally. Physicians remain active participants in every important clinical decision while continually extending their own thinking through structured interaction with intelligent systems. The technology becomes a catalyst for cognitive development rather than cognitive substitution.

The implications extend well beyond diagnosis. Medical education should prepare future physicians not simply to use AI tools but to collaborate with them effectively. That requires learning how to formulate better questions, recognize when AI reasoning conflicts with clinical context, identify hidden assumptions, and integrate machine-generated insights without surrendering professional judgment. These are becoming core competencies of modern medical practice.

Health care organizations should also reconsider how success is measured. The objective should not be the number of decisions transferred from physicians to AI. It should be whether physician-AI collaboration consistently produces better clinical judgment, better patient outcomes, and better learning than either could achieve alone.

Medicine has reached similar crossroads before. Laboratory medicine, advanced imaging, and molecular diagnostics all expanded the physician’s ability to understand disease. None attempted to replace clinical judgment. They strengthened it. Artificial intelligence presents an opportunity to do the same, but only if we choose to pursue that path deliberately.

Medicine does not need a future in which physicians compete with AI. It needs a future in which physicians and AI learn to think together. That future will not emerge on its own. It must be intentionally designed.

The question before us is not whether AI will become more intelligent. It almost certainly will. The more important question is whether we will use that intelligence to replace human clinical judgment or to elevate it. I believe the future of medicine depends on making the right choice.

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.