The article reviews literature relating the application of artificial intelligence to emotion recognition in children and adolescents and its potential clinical applications. For the review, articles from the following databases were used: PubMed, Scopus, CINAHL Ultimate, MEDLINE Ultimate, IEEE Xplore and Wiley Online Library. Out of the 136 results yielded, 23 studies were selected for review. The studies used a variety of algorithms, AIs, and machine learning programs with varying populations and goals in mind. These were split into eight sections based on their applications: pediatric mental health assessment, school and community recognition, behavioural interventions, social robots for children with Autism Spectrum Disorder (ASD), emotion recognition in ASD, virtual environments and drawings in emotion recognition, serious games for emotional development, and affective robots for hearing impaired children. Furthermore, the AI tool that each study focused on was put in a matrix of four functions: Emotional Assessment, Therapeutic Intervention, Clinical Decision Support, and Communication Support. The matrix conveys the variability in applications of the specific AI tools used across studies. Overall, the article discusses how AI-powered devices have improved in capturing children’s emotional experiences, extending communication possibilities for nonverbal children, and making therapeutic processes more engaging. However, models still need more nuanced recognition of psychological distress and negative emotions and should be understood as a complement to health care rather than a replacement for human empathy.