A guest essay in the New York Times prompted Dr. Jayne, HIStalk's physician columnist, to examine whether AI scribes are actually delivering on their promise of reducing clinician burden. Her conclusion is cautious: the tools may be trading one kind of work for another. Physicians still need to review, correct, and take responsibility for AI-generated notes — a process that carries its own cognitive overhead and medicolegal weight, even if keystrokes go down.
The piece lands at a moment when AI documentation tools are being adopted rapidly across health systems, often driven by administrator enthusiasm for efficiency gains rather than clinician input. Dr. Jayne notes that rollout decisions frequently outpace evidence, and that the clinicians doing the reviewing are rarely the ones who shaped the procurement or implementation.
For HTM and clinical informatics teams supporting these deployments, the implication is practical: ambient AI tools are not set-and-forget. Ongoing attention to how notes are being generated, reviewed, and corrected — and whether clinicians have a clear feedback channel — matters as much as the initial integration work.
Source: HIStalk