Towards safe conversational agents in healthcare.

The public is increasingly using conversational assistants like Siri, Alexa, and Google Assistant to find medical advice and self-diagnose. This narrative review summarizes three facets of safety: system (data privacy/security), patient (risks of acting on inaccurate information), and perceived (patient trust in the system). Future research should address all three safety facets, and the results should be transparent to consumers.