AI literacy in undergraduate medical education: a competency-based interpretive framework for curriculum and assessment

Abstract

Artificial intelligence (AI) is becoming a core educational concern in undergraduate medical education as AI-enabled tools increasingly shape clinical workflows, learning environments, and patient care. The challenge is no longer simply whether AI should be included in the curriculum, but how AI literacy should be bounded for undergraduate learners and translated into teachable, observable, and assessable educational outcomes. This focused conceptual narrative review synthesized literature on AI literacy and related constructs in undergraduate medical education, using a structured search and interpretive synthesis with competency-based medical education (CBME) as an interpretive lens. PubMed and ERIC were searched for English-language literature from 1 January 2020 to 15 April 2026. Local screening records identified 94 standardized bibliography records, 66 records screened after deduplication, 40 full-text reports assessed, and 30 publications contributing to the final synthesis. Five recurring domains were identified: Foundational AI knowledge; applied clinical interpretation and use; data literacy and critical appraisal; ethics, law, and professional responsibility; and human–AI collaboration and professional formation. Through a CBME lens, these domains can be translated into learning outcomes, contextualized tasks, observable performances, and programmatic assessment evidence. The literature most strongly supports conceptual clarification, domain identification, and curricular translation, whereas evidence for longitudinal development, observable performance, and validated undergraduate assessment remains limited. The proposed framework, examples, milestones, and rubric anchors are synthesis-informed design propositions that require empirical validation before high-stakes use.