Medical students and clinical teachers experiences and perceptions of using AI-based simulated patients in psychiatry clinical skills training, a qualitative analysis

In psychiatry clinical skills training at Kent and Medway Medical School, UK, 16 students and nine teachers viewed an LLM-based simulated patient as a flexible supplement, not a replacement, despite realism and delay barriers.

• Why it matters: The study addresses uncertainty about how AI-based simulated patients are experienced and governed in psychiatry training, including whether they can support communication practice without undermining ethics or human interaction.
• What they did: Participants used an LLM-ASP tool (ChatGPT Advanced Voice Mode, Streamlit, LangChain) for about 10 minutes before focus groups, then audio-recorded discussions were analysed via reflexive thematic analysis with multi-cycle coding.
• The result: Both groups identified themes around educational value and ethical, cultural, and governance issues, with shared concerns about safeguarding, data storage, consent, demographic bias, and over-reliance, while barriers included limited emotional realism, robotic delivery, and response delays.

AI summary of the abstract · may contain errors · not medical advice · read the paper.