Adverse effects, safety concerns, and harms of mental health chatbots for adolescents: A scoping review

This scoping review will systematically map adverse effects, safety concerns, and documented harms of mental health chatbots for adolescents aged 12-18. The review addresses a critical gap in the literature: while 90% of mental health chatbot studies report efficacy outcomes, few systematically document safety outcomes despite their clinical importance. Following Arksey and O'Malley's methodological framework and PRISMA-ScR guidelines, we will search PubMed, PsycINFO, Web of Science, PubMed Central, and grey literature from 2015-present. The review will include randomised controlled trials, observational studies, qualitative studies, mixed methods studies, case reports, and evaluation studies that report (or fail to report) adverse effects, safety incidents, privacy concerns, or user experience harms. Primary outcomes include clinical harms (symptom worsening, dependency, reinforcement of maladaptive beliefs), safety incidents (inappropriate responses, crisis mismanagement), privacy concerns (data breaches, confidentiality violations), and user experience harms. The review will also document patterns in safety reporting by funding source. Expected outcomes include a comprehensive map of documented harms, identification of gaps in safety reporting, and recommendations for future research and clinical practice. This work will inform evidence-based guidance for clinicians