The potential impact of generative AI across Miller's pyramid of clinical competence: A systematic review

This systematic review examines the effectiveness and limitations of generative artificial intelligence (GenAI) in supporting the development of clinical competence in nursing education across Miller's pyramid, synthesizing evidence to inform the design of educator-guided, Expert-in-the-Loop hybrid educational frameworks.

Nursing education faces challenges from ongoing pedagogical innovations linked to GenAI. This technology offers transformative potential for personalized learning; however, its rapid integration into curricula lacks a systematic evidence base.

This systematic review used seven databases (CINAHL, PubMed, Scopus, Web of Science, Embase, PsycINFO and Cochrane) to find studies published from November 2022 to May 2025.

Studies were screened using predefined criteria. Methodological quality was assessed with the Mixed Methods Appraisal Tool. A narrative synthesis was conducted, interpreting the findings based on Miller's pyramid of clinical competence.

Twelve studies met the inclusion criteria. GenAI applications were concentrated at the "Knows" (knowledge) and "Knows How" (application) levels of Miller's pyramid, with no studies assessing real-world practice ("Does" level). While GenAI models passed licensure exams and improved cognitive and affective outcomes, they also had limitations, including factual inaccuracies, failures in clinical judgment and performance disparities.

GenAI is a powerful yet error-prone tool that necessitates systematic pedagogical redesign, not merely technological adoption. The Expert-in-the-Loop model, where educators validate generative AI outputs, may be key to transformations in nursing education. Embedding generative AI in educator-led frameworks can cultivate nursing professionals who leverage technology critically and ethically, ensuring care is both competent and human-centered. Registration This study was registered on PROSPERO (ID: CRD420251060512).