Artificial intelligence as a constructivist scaffold in anaesthesia education: a mini review

Artificial intelligence (AI), including generative AI and large language models, is increasingly used in health professions education, but evidence for educational benefit in anaesthesia remains limited and heterogeneous. This focused Mini Review incorporates literature identified through August 2026 and distinguishes anaesthesia-specific empirical evidence from feasibility studies, extrapolation from broader health professions education, and proposed applications. Using constructivist learning theory, we conceptualise AI not as an autonomous teacher or objective assessor, but as a contingent and temporary scaffold that should be adapted to learner performance, progressively faded, and followed by assessment of independent transfer. We further integrate cognitive apprenticeship through modelling, coaching, articulation, reflection, and exploration while maintaining the essential role of human supervisors. Available anaesthesia-specific studies involving AI-assisted regional anaesthesia training, ultrasound interpretation, virtual patient encounters, and simulation analysis demonstrate feasibility or task-specific benefits but do not yet establish improvements in clinical reasoning, debriefing, competence, or entrustment. Building on recent reviews of AI in anaesthesiology education, this review operationalises contingency, fading, transfer, cognitive apprenticeship, and assessment safeguards for anaesthesia training. AI-generated feedback, portfolio summaries, and scores should be regarded as fallible decision-support and should not independently determine summative assessment, progression, remediation, or entrustment. Future research should evaluate whether AI provides added value beyond established simulation and human-led supervision, focusing on reasoning, retention, transfer, safe escalation, and equity.