Physician Associate Student Use of Large Language Models to Support Learning: A Phenomenological Study

As generative artificial intelligence (AI) tools become increasingly prevalent in higher education, their integration into physician associate (PA) education remains underexplored. The existing literature primarily focuses on short-term outcomes, with limited attention to learner engagement, critical thinking, or institutional support.

This qualitative study employed semistructured interviews with 8 self-identified AI-using didactic PA students from one private and one public institution. Interpretative phenomenological analysis and constructivist learning theory guided the analysis. Transcripts were thematically coded using an a priori codebook and refined iteratively.

Three key themes emerged: (1) AI integration in learning practices: students used AI for generating practice questions, clarifying content, and rehearsing for clinical assessments; (2) student perceptions and attitudes: AI was seen as efficient and supportive, though caution was exercised due to inaccuracies and perceived limitations; (3) institutional context and support: formal guidance was minimal, with most students navigating AI use independently in permissive but understructured environments.

Didactic PA students interviewed for this study actively incorporate generative AI into their learning through trial, reflection, and adaptation. Despite limited institutional direction, participants demonstrated emerging digital literacy and critical engagement. These findings underscore the need for AI-integrated curricular frameworks and faculty development to support the ethical and pedagogically sound use of AI by students. Programs that partner with learners in this evolving landscape may better prepare future clinicians for technology-enhanced practice.