From Clinical to Educational Data: Ethical Pathways for Re-purposing AI-Generated Patterns in Medical Training

Artificial intelligence is increasingly used in healthcare to generate patterns, classifications, and predictive outputs from clinical data. A growing number of these outputs are later re-purposed for medical education through simulations, synthetic cases, case-based learning, and other pedagogical resources. This shift raises an underexplored ethical problem: once AI-derived clinical outputs are moved from care settings to training environments, their function changes from decision support to epistemic and formative mediation. This study examines that transition and argues that AI-derived clinical patterns should be understood as epistemic artefacts capable of shaping how future clinicians classify disease, interpret uncertainty, and develop professional judgment. Methodologically, the study adopts a conceptual and normative approach grounded in the ethics of artificial intelligence, biomedical ethics, and educational governance. The study develops an ethical trajectory model structured around four stages: patient data, AI-derived pattern, pedagogical artefact, and educational impact. On this basis, it distinguishes clinical AI ethics from educational reuse ethics and identifies a set of governance implications for the responsible re-use of AI-derived clinical knowledge in medical training. The study contributes a conceptual framework for assessing secondary educational uses of AI-generated clinical outputs and highlights the need for future empirical work on pedagogical design, documentation practices, and ethical review criteria in medical education.