Beyond the Hype: Artificial Intelligence as Decision Support in Emergency and Prehospital Medicine—A Critical Narrative Review

Artificial intelligence (AI) and machine learning (ML) are increasingly being explored in tertiary emergency departments and prehospital emergency medical services because these settings combine time-critical decisions, incomplete information, and persistent operational strain. This critical narrative review examines where current evidence is strongest, where it remains limited, and what ethical and practical conditions are required for responsible adoption. The literature suggests that the most credible near-term applications are bounded forms of decision support, particularly in triage and early risk stratification, selected time-sensitive diagnostic pathways such as stroke, operational forecasting, dispatch support, destination decisions, and supervised documentation assistance. However, the field remains constrained by retrospective study designs, heterogeneous endpoints, limited external validation, uncertain transportability, uneven subgroup performance, and a persistent gap between model performance and real-world clinical usefulness. In emergency and prehospital care, these limitations matter because bias, overreliance, and poor workflow integration can rapidly translate into patient harm or operational disruption. Current evidence supports cautious optimism rather than technological inevitability. The most defensible role for AI in acute care is not the replacement of clinicians or EMS professionals, but carefully governed augmentation of human judgment through locally validated, transparent, and continuously monitored decision-support tools.