Artificial intelligence in pediatric pain: a systematic review

BACKGROUND: Pain is a challenging, multifaceted symptom reported by most pediatric patients. This systematic review aims to explore the progress and effectiveness of applying artificial intelligence (AI) technology in pediatric pain management. METHODS: A comprehensive search of PubMed, Embase, Web of Science, Cochrane, Scopus, IEEE Xplore, ACM Library, and ClinicalTrials.gov was conducted. The search combined pain-related terms ("Pain management", "Pain assessment", "Pain measurement", "Pain relief", "Pain control", "Analgesics", "Pediatric pain"), age-related terms ("Children", "pediatrics", "Neonate", "Infant"), and AI core sub-domain terms ("Artificial intelligence", "Machine learning", "Deep learning", "Convolutional neural network", "Support vector machine", "Random forest", "Long short-term memory") using strict Boolean operators (AND/OR/NOT) published up to March 20, 2026. AI models, their validation, and performance were summarized. RESULTS: The analysis of 71 studies revealed distinct AI application patterns in pediatric pain management. Fifty-nine studies focused on pain assessment using deep learning (post-2020: 86.7%) and classical machine learning (pre-2015: 83.3%) through facial expression analysis (40.8%), multimodal fusion (25.4%), or physiological signals (9.9%). Most employed observational designs (57.7%) with small (