Artificial intelligence in physical examination teaching in Latin America: a critical narrative review and conceptual model proposal

Artificial intelligence (AI) is reshaping medical education, particularly in the teaching of physical examination and the development of clinical judgement in digitally mediated contexts. This study presents a critical narrative review examining the ethical, pedagogical, and humanistic implications of AI integration into physical examination training in Latin America. A structured search of literature published between 2018 and 2025 was conducted across PubMed, Scopus, Web of Science, SciELO, and Google Scholar. Thirty-one peer-reviewed studies and three institutional documents met predefined relevance criteria and were analyzed through thematic synthesis. Four thematic domains emerged: (1) AI-assisted clinical simulation and automated feedback, (2) curricular integration and institutional implementation strategies, (3) governance and ethical supervision frameworks, and (4) emerging challenges related to digital literacy, technological dependence, and preservation of clinical judgement. Evidence suggests that AI enhances procedural precision and formative feedback; however, its educational value remains complementary and dependent on structured human-in-the-loop supervision. Based on these findings, the Modelo Educativo Digital basado en Inteligencia Artificial (or Medical Education with Artificial Intelligence) (MED-IA) conceptual model is proposed, framing clinical competence development across three interconnected levels: technical execution, experiential patient interaction, and reflective judgement. The model integrates technological mediation with ethical oversight and humanistic formation. These findings highlight the need for transparent governance frameworks, teacher digital literacy, and context-sensitive institutional policies to ensure responsible AI implementation in Latin American medical education.