The widespread adoption of AI in medicine may increase the number of health care professionals rather than reduce them, according to a new analysis published Sept. 12 in The New England Journal of Medicine. The finding challenges predictions that AI agents will displace clinicians across specialties like radiology, primary care, and psychiatry.
Dr. Dhruv Khullar, an associate professor of population health sciences at Weill Cornell Medicine and a hospitalist at NewYork-Presbyterian/Weill Cornell Medical Center, examined economic theory and the history of technology in health care to build his case. He said that while some jobs will change or disappear, the long-term trajectory points toward workforce expansion.
"Economic theory and history offer distinct reasons to believe that in the long run, adoption of AI agents-even models capable of performing some forms of cognitive work-could lead to an expansion, rather than a contraction, of the clinical workforce," Dr. Khullar said.
The Jevons paradox in health care
Dr. Khullar pointed to cataract surgery and joint replacement as examples of Jevons paradox. Those procedures became more efficient over time, requiring less clinical effort and shorter recovery periods while improving safety. The result was not fewer surgeons performing fewer operations. More patients received treatment.
AI could trigger a similar dynamic, especially if systems are priced near marginal cost and drive down the overall expense of delivering care. Lower costs typically unlock demand that was previously suppressed by price or access barriers.
Why automating tasks is not automating jobs
The "lump of labor" fallacy assumes a fixed amount of work exists. Dr. Khullar argued that work changes in response to technology. AI may let clinicians prevent or treat conditions in ways that are not currently possible. "AI may increase the demand for new professional capabilities by creating new treatments, care modes, and opportunities for specialization," he wrote.
Medicine depends on interconnected steps - interpreting tests, developing treatment plans, negotiating with patients. Drawing on O-ring theory, which originated from the Space Shuttle Challenger disaster, Dr. Khullar noted that a single error at any step can undermine outcomes. High-stakes clinical situations will continue to require human supervision for safety and trust. When AI automates some tasks, the value of the remaining human tasks may rise. "In fact, if AI automates some clinical tasks, the value of nonautomated, human tasks may increase," he said.
Why this matters for health care professionals
For clinicians worried about job displacement, the analysis offers a counter-narrative grounded in economic history. The immediate shift involves roles changing rather than vanishing. Specialties that appear highly exposed to AI Agents & Automation may see demand shift toward tasks that require judgment, empathy, and complex coordination - skills AI cannot replicate.
Health care organizations planning workforce strategy should watch how payment models adapt. If AI for Healthcare lowers the cost of common services, volume could increase enough to sustain or grow clinical staffing. The key variable is not what AI can do, but how the system prices and deploys it.
Your membership also unlocks:
