AI-Enabled Physical Activity Intervention Strategies for Children and Adolescents: A Scoping Review.

BACKGROUND/OBJECTIVES: Most children and adolescents do not meet recommended physical activity (PA) levels. Artificial intelligence (AI) may enable personalized and adaptive interventions, but evidence across delivery formats remains unclear. This scoping review mapped AI-enabled PA interventions for children and adolescents aged 6-18 years. METHODS: Following PRISMA-ScR and Joanna Briggs Institute guidance, six databases were searched on 24 May 2026 for English-language studies published since 2018. Eligible studies used AI in intervention delivery and empirically evaluated at least one participant-level PA-related outcome following exposure to the intervention. The review protocol was registered during data analysis. RESULTS: Seven studies published between 2022 and 2026 were included. Primary AI techniques were computer vision (n = 3), generative AI or large language models (n = 2), reinforcement learning (n = 1), and evolutionary computation (n = 1). Interventions included immersive systems (n = 2), conversational agents (n = 2), exergames or serious games (n = 2), and a robot (n = 1). Three studies used controlled designs and four were single-arm pilot or feasibility studies. CONCLUSIONS: AI supported movement recognition, content generation, personalization, and adaptive feedback. However, the small, heterogeneous, and predominantly early-stage evidence base precludes conclusions about effectiveness. Rigorous controlled studies and transparent AI-specific reporting are needed.