Teaching AI-Informed Diagnostic Reasoning in Dental Education: A Conceptual Framework

Artificial intelligence (AI) is increasingly present in dental education and clinical practice through systems that support image interpretation, lesion detection, risk estimation, treatment planning, documentation, and access to clinical information. Existing curriculum frameworks appropriately emphasize foundational knowledge, ethical use, governance, and human-centered practice. Broad AI literacy, however, does not by itself teach students how to reason when an algorithmic output influences diagnosis or management. This perspective proposes a conceptual framework for teaching AI-informed diagnostic reasoning in dental education. The framework positions AI as a contributory input rather than an independent diagnostic authority and organizes learner performance around three linked tasks: interpreting the output, reconciling it with patient-specific evidence, and justifying an accountable management decision. It is conceived as a metacognitive overlay on existing clinical reasoning competencies rather than as a standalone clinical discipline. The framework can be operationalized in Objective Structured Clinical Examinations (OSCEs) through staged disclosure of AI output, independent learner interpretation before disclosure, structured revision of reasoning after disclosure, and a developmental rubric that combines domain scores with a global judgment. Worked endodontic and caries-risk vignettes illustrate assessment of reasoning quality rather than agreement with AI. The distinct educational challenge introduced by AI is not simply the arrival of additional information. Algorithmic outputs may carry an authority effect, obscure uncertainty and data provenance, encourage automation bias, and reproduce limitations or inequities in training data. Learners must therefore evaluate both the clinical meaning of the output and the conditions under which it was generated. The framework is intended for curriculum development and formative assessment and provides a basis for subsequent study of feasibility, faculty calibration, scoring reliability, validity, learner acceptability, and educational impact.