The AI Revolution in Pediatric Emergency Medicine: Transforming Care in the Golden Hour

Abstract

Artificial intelligence (AI) is rapidly transforming in clinical practice, but its impact is especially profound in paediatric emergency medicine, where a few minutes can influence an entire life. Building on the work summarized in “The AI Revolution in Medicine” and related studies on digital phenotyping and autonomous diagnostic tools, this article reviews how AI is beginning to change pediatric emergency care: from pre-arrival assessment and triage to diagnostic imaging, deterioration prediction, and decision support.¹
We summarize emerging evidence on triage and risk stratification models, sepsis and clinical deterioration prediction systems, AI-assisted diagnostic imaging, and the first large-scale language model  applications in pediatric emergency departments (PEDs). Machine learning tools have been shown to outperform traditional triage scales in identifying critically ill children, improving the early detection of sepsis, and supporting the interpretation of pediatric images, although much of the work remains retrospective and single-center. At the same time, pediatric evidence is scarcer than in adults, and issues of data shortages, bias, workflow integration, and ethics are particularly sensitive when dealing with children.