Popular AI Models Get Medication Doses Wrong, Some Dangerously: Doximity (2026 Update)
OVERVIEWResearch from Doximity reveals that popular general-purpose AI models—including ChatGPT, Gemini, and Claude—make dangerous medication dosing errors. The study demonstrates that specialized clinical AI tools significantly outperform these general models on dose verification tasks, highlighting a critical safety gap as health systems adopt off-the-shelf LLMs for clinical decision support. The findings show error rates that range from concerning to potentially lethal, with some recommendations falling outside safe dosing ranges by orders of magnitude. Doximity's specialized dose verification tool achieved substantially higher accuracy, pointing to the value of domain-specific training and clinical validation. This research underscores a governance challenge: general-purpose AI models are being deployed in clinical contexts despite lacking the specialized training and validation required for medication safety decisions.
Subscribe to read full wiki analysis
Full wiki content — including source provenance, contradiction tracking, related pages, and analyst notes — is available to Subscriber members.
View Plans & SubscribeNew users get a 7-day trial — no credit card required.