Beyond Borders: Global Mental Health Conference

The United States is facing a serious youth mental health crisis. Approximately 18% of adolescents between the ages of 12 and 17 experience depressive episodes each year, and nearly 40% of those receive no mental health care (McBain et al., 2025). Into this gap, generative AI chatbots have become informal mental health resources. A 2025 nationally representative survey of 1,058 youth aged 12 to 21 found that one in eight adolescents use AI tools such as ChatGPT for mental health counsel, without professional support (McBain et al., 2025).
Despite the scale of this phenomenon, its risks remain critically understudied. Ng (2025) found that generative AI promotes emotional attachments in vulnerable youth and risks reinforcing pathology. Sharma et al. (2025) revealed that AI research in adolescent mental health has focused heavily on diagnostic tools, leaving unstructured chatbot interactions under-investigated. Feng et al. (2025) examined AI chatbot effectiveness in structured clinical interventions but did not address the self-directed, unsupervised way most teens engage with these tools. Scholarship on AI-generated health misinformation has similarly concentrated on COVID-19 and vaccination, leaving mental health almost entirely unexamined (Suarez-Lledo & Alvarez-Galvez, 2021).
I present a conceptual framework examining adolescent self-diagnosis of depression via AI chatbots and the misinformation risks embedded in that process. I ask three questions: whether AI-generated responses to depressive symptom disclosures are clinically accurate, whether teens act on those responses in ways that delay formal help-seeking, and whether vulnerable subgroups, including LGBTQ+ youth, rural adolescents, and those from non-English-speaking populations, face disproportionate misinformation exposure (Grenon et al., 2025). I call for interdisciplinary collaboration across clinical psychology, digital health, and communication science to evaluate AI depression guidance, develop adolescent-specific AI literacy interventions, and establish policy standards before resiliency efforts are undermined by the very technologies young people are already turning to for help.