Answers are grounded in this paper and cite their sources.
Residents increasingly ask large language models to name a lesion before diagnosing it themselves. The practice contributes a specific harm: never‐skilling, a phenomenon in which reasoning is never acquired, and is distinct from 'deskilling' or 'model error'. Forming a differential first is necessary but insufficient. Departments should require documented differentials, supervise disagreement, and examine unassisted reasoning.
No takes yet. Share an insight, caveat, or question.
Apurv Khandelwal (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: