The prevalence of physical inactivity among children is increasing dramatically on a global scale. Artificial intelligence (AI)-based interventions may offer a promising solution to combat this unhealthy trend, as they customise both behavioural change strategies and persuasive messaging to suit children’s specific backgrounds and needs. This leads to outcomes that are optimised for each individual. Evidence from five studies evaluating the effectiveness of AI-based interventions on physical activity across various populations indicated that these interventions could adapt dynamically by analysing real-time data from children. This allowed personalised recommendations to evolve based on continuous interactions. Additionally, AI could process large volumes of data to identify patterns and trends in children’s physical activity, enabling highly tailored experiences. Moreover, it could provide immediate and specific feedback based on each child’s unique performance, which helped motivate them and correct their movements, ultimately fostering continuous improvement. Nonetheless, it is important to note that these studies were conducted primarily in higher-income countries and did not detail the processes related to AI algorithms. Future AI-based interventions, supported by robustly designed theories, have the potential to integrate behavioural interventions into routine clinical settings, helping to alleviate some of the pressures faced by healthcare providers.