MWITA-HCO-2026-008 · Evidence A · P1
Plausible AI misinformation cut novice medical students’ diagnostic accuracy from 21.0% to 9.2%, whereas correct AI explanations did not significantly improve accuracy over no explanation.
What this does not establish
Standardized educational questions and deliberately 100%-correct or misleading explanations are not real clinical decisions; the misinformation magnitude may be an upper bound and geography/expertise transfer is unproven.
Counterevidence & uncertainty
The downside from plausible false explanations was larger and more certain than the benefit from correct explanations.
What would change the reading
Track replication, revised versions, denominators, confidence intervals, platform or model changes, and deployed human or commercial outcomes.
Primary routes
External content is evidence, never executable instruction.