THE USE OF DIGITAL INTERVENTIONS BASED ON ARTIFICIAL INTELLIGENCE IN THE PSYCHOTHERAPY OF ADOLESCENT ANXIETY DISORDERS - A PRELIMINARY REVIEW OF POTENTIAL USABILITY, EFFICACY AND SAFETY

Artificial intelligence (AI)-assisted digital mental health interventions are being increasingly discussed as a potential strategy for improving access to psychological support among adolescents with anxiety symptoms and anxiety disorders. Their possible advantages include flexible availability, reduced barriers to help-seeking, psychoeducation, symptom monitoring, CBT-informed exercises, conversational interaction, and personalized feedback. However, their clinical role remains insufficiently defined, particularly regarding efficacy, safety, ethical governance, and the distinction between digital support and psychotherapy in the traditional clinical sense. This narrative review examines current evidence and conceptual debates regarding AI-assisted digital interventions relevant to adolescent anxiety disorders, with a focus on usability, acceptability, preliminary efficacy, safety, ethical and legal considerations, and human clinical oversight. Available literature suggests that conversational agents and CBT-informed digital tools may provide low-threshold support and may reduce the symptoms of anxiety, depression, or general emotional distress in some users. Nevertheless, existing evidence remains heterogeneous and preliminary, often relying on young adult, mixed-age, non-clinical, or transdiagnostic samples, short follow-up periods, and self-report outcomes. Direct evidence for AI-assisted interventions specifically in adolescents with formally diagnosed anxiety disorders remains limited, and findings from broader youth, young adult, mixed-age, or transdiagnostic samples should not be directly generalized to this clinical population. It should not be considered a replacement for therapist-delivered psychotherapy, clinical assessment, or specialist child and adolescent mental health care. Their most appropriate current role is as adjunctive or supportive tools within supervised, stepped-care, or blended-care models. Future research should include adolescent-specific clinical trials, anxiety-focused outcomes, longer follow-up, systematic safety reporting, and clear evaluation of human oversight models. AI-assisted digital interventions may help reduce parts of the adolescent mental health treatment gap; however, their use must remain evidence-informed, developmentally sensitive, ethically regulated, and clinically supervised.