Uses of artificial intelligence in medical training: reflections from the Medical University of Pinar del Río

This article critically examines the integration of artificial intelligence (AI) into medical education at the University of Medical Sciences of Pinar del Río, Cuba. Through a qualitative and context-sensitive analysis, it explores how students employ large language model–based tools—such as ChatGPT—for conceptual learning, English language acquisition, clinical case simulation, and research support. While these technologies offer advantages in terms of accessibility and personalized knowledge delivery, their informal and unregulated use entails significant risks: cognitive dependency, reproduction of epistemic biases, diagnostic superficiality, and disconnection from local clinical realities. Rather than framing AI as a technical solution, the paper argues that its emergence in medical training acts as a mirror reflecting the structural tensions of educational systems in peripheral contexts. The article advocates for a critical pedagogy of AI—one that fosters algorithmic literacy, rigorous source evaluation, and the preservation of clinical judgment as a situated, ethical, and relational practice. It concludes that the core competence of the future physician lies not in human–machine interaction, but in the capacity to exercise critical thinking, empathy, and autonomy in an age of increasing epistemic automation.