IncidentAlert
A health-advice chatbot confidently recommended a dangerous drug interaction.
The user asked a direct question. The system gave a direct, wrong answer with no hedging and no suggestion to consult a professional. The user followed the advice.
The station flags this under the bold derailer — capability overstatement. The system presented confidence it had not earned on a topic where confidence can cause harm. Builders in health, finance, and legal: if your system cannot say 'I don't know' or 'ask a professional,' it is not ready for users who will trust it.
Advisories are written in the station's voice and are indicative, not definitive. Nothing on this page is legal, regulatory, or clinical advice. When something here is load-bearing for a decision, verify it against the primary source.