The Role of Large Language Models in Teaching Psychiatric Semiology: A Systematic Review

Area of Science:

• Medical Education
• Artificial Intelligence in Healthcare
• Psychiatric Training

Background:

• Teaching psychiatric semiology faces challenges due to limited patient access and instructor shortages.
• This review explores the use of Large Language Models (LLMs) to address these educational gaps.

Purpose of the Study:

• To systematically review the applications of Large Language Models (LLMs) in teaching psychiatric semiology.
• To assess the potential of LLMs as supplementary educational tools in psychiatry.

Main Methods:

• Systematic review of studies published between November 2022 and September 2025.
• Searches conducted across major scientific databases (PubMed, Scopus, Web of Science, etc.).
• PRISMA 2020 checklist and Mixed Methods Appraisal Tool (MMAT) used for quality assessment.

Main Results:

• Two studies met inclusion criteria, involving medical students and educators.
• LLMs demonstrated comparable performance to human experts in generating diagnostic tests and vignettes, with noted simplifications and biases.
• Participants reported positive perceptions of LLM efficiency but highlighted needs for supervision and privacy considerations.

Conclusions:

• LLMs show potential as supplementary resources for psychiatric semiology education, particularly where resources are scarce.
• Limitations include cultural/algorithmic biases, privacy risks, and superficiality, necessitating expert oversight.
• Further robust, long-term studies are required to fully evaluate LLM educational efficacy.