Chatbots for Children and Adolescents: A Meta-analysis of Chatbots for Mental Health Issues and Psychosocial Well-Being Outcomes

Children and adolescents experience rapid developmental changes that shape both the opportunities and challenges of using chatbots for mental health support. Younger children are still developing emotion regulation and social learning skills, while adolescents navigate identity formation and autonomy. These transitions influence how youth understand, trust, and engage with chatbots. Existing reviews on mental health chatbots tend to focus on young adults, leaving little evidence for younger users. This pre-registered meta-analysis extends existing reviews by synthesizing experimental evidence focused on youth below 18 years old. Through seven databases and backward citation tracking searched between January, 2014 and March, 2025, 2,398 studies were retrieved. After double-blinded screening, 12 studies were included (Range(age) = 3 similar to 18 years old, N-total = 2,961, N-ES = 51). Meta-analysis results showed that chatbots are effective for reducing mental health issues (ES = 0.47, p = .027), but no evidence for improvement in psychosocial well-being outcomes. Moderator analysis showed that, compared to text-based chatbots, audio-based ones are more effective for children and adolescents. These findings advance developmental science by specifying which design features enhance mental health chatbots' effectiveness, providing guidance for creating digital tools that align with children's and adolescents' developmental needs. Lastly, readers must interpret the results with some key limitations in mind, including the small sample size, treatments' short durations, significant heterogeneity, self-reported outcome measures, and publication bias.