Evaluation of artificial intelligence-generated vignettes depicting patient chatbot use in psychiatric contexts.
Journal:
NPJ digital medicine
Published Date:
Apr 7, 2026
Abstract
Chatbot-related mental health risks have attracted increasing public attention, yet educational case materials for clinicians remain limited. We evaluated ChatGPT-5 Pro's ability to generate psychiatric vignettes involving patient chatbot use. Three board-certified psychiatrists rated the vignettes on chatbot relevance, diagnostic sufficiency, explanation quality, and safety. While chatbot relevance and diagnostic sufficiency were high, safety ratings were lower, underscoring the need for expert review before educational use.
Authors
Keywords
No keywords available for this article.