AI chatbots and youth suicide risk: current evidence, critical gaps, and a clinical research agenda

Abstract Generative AI chatbots are increasingly used by youth for mental health support, but their safety for managing youth suicide risk remains poorly understood. In this paper, we review research on how general-purpose chatbots and AI companions detect and respond to suicide risk disclosures, finding that existing studies rely almost entirely on single-turn, researcher-generated prompts that are not designed with youth in mind. Chatbot performance degrades substantially with indirect, ambiguous, or gradually disclosed risk, which are patterns characteristic of adolescent communication, and the real-world impact of chatbot interactions on youth suicide risk remains unknown. We propose a youth-focused clinical research agenda to address these gaps, organized around four priorities: understanding and integrating youth perspectives through qualitative and participatory research; improving youth-specific risk detection and evaluation methods; examining real-world impacts of chatbot use among youth through prospective outcome research; and improving connection of at-risk youth to crisis resources and trusted adults. Progress will require multidisciplinary collaboration and must evolve alongside rapidly changing chatbot capabilities.