JMIR Publications

ABSTRACT

Artificial intelligence (AI) is rapidly reshaping medical education and challenges the validity of traditional knowledge‑centric assessments. Drawing on the principles of andragogy, this article argues that evaluation in the AI era should be anchored in adult learners’ characteristics, experiences, needs, and psychological safety rather than in decontextualized recall. Personal and situational characteristics, prior experience, evolving needs and aspirations, inner motivation, concerns and vulnerability, learning processes, and well‑known limitations of self‑assessment are the main andragogic parameters used as a framework to propose mainly non‑AI or minimally AI‑dependent assessment tools. These include simulation‑based and workplace‑based assessments, objective structured clinical examinations, structured oral exams, project‑based tasks, multi‑source feedback, reflective portfolios, and real‑time performance prompts that evaluate clinical reasoning, communication, professionalism, and pattern recognition. The article also outlines formative uses of AI outputs as raw material for critique, bias detection, and evidence checking, while reserving high‑stakes decisions for assessments of live performance in authentic contexts. Taken together, these strategies shift assessment from simple AI‑susceptible assessments toward complex, competency‑based, human‑centered assessment that supports learners’ autonomy, intrinsic motivation, and responsible engagement with AI in clinical practice.