Opinion: AI won’t enhance physician autonomy. It will further diminish it
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- The AMA's 2026 survey found 81% of physicians already use AI professionally — more than double the rate three years ago — while 85% say they want a real voice in how it gets adopted in their own practices, a gap Hardin frames as adoption outracing consent.
- A JAMA study published in April reported ambient AI scribes saved about 16 minutes of documentation time per eight hours of patient care, or roughly two minutes per hour of patient care.
- The FDA in January narrowed the definition of what counts as a regulated device in clinical decision support, leaving a wide swath of AI tools with less federal oversight as long as the clinician independently reviews the recommendation.
- Hardin argues the FDA's narrowed oversight means decreased scrutiny of the AI tool combined with increased liability for the physician clicking "accept," following a long pattern of risk being shouldered by those with the least power to resist.
- The opinion piece draws parallels to prior physician disempowerment — The Match for residency placement, productivity metrics for visit length, RVU models for time valuation, and prior authorizations for treatment choice — before AI reached the zone of clinical judgment itself.
Why it matters: The AMA's split between 81% adoption and 85% wanting input shows consent is lagging deployment, and the FDA's January narrowing of clinical decision support oversight means that when an AI-generated recommendation deviates from the standard of care, the physician at the top of the chart — not the trillion-dollar AI company — absorbs the liability.
Ask SkimNews
