"If This Person is Suicidal, What Do I Do?": Designing Computational Approaches to Help Online Volunteers Respond to Suicidality

ACM CHI 2024 (To Appear)
Logan Steplon, Sunniva Liu, Cindy Liu, Irene Hong, Robert Kraut, Haiyi Zhu

Online platforms provide support for many kinds of distress, including suicidal thoughts and behaviors. However, because many platforms restrict suicidal talk, volunteer on these platforms struggle with how to help suicidal people who come for support. We interviewed 11 volunteer counselors in a large online support platform, including after they role-played conversations with varying severities of suicidality, to explore practices and challenges when identifying and responding to suicidality. We then presented Speed Dating design concepts around suicide detection, training, real-time guidance, and emotional support. Participants wanted support from suicide-trained staff and emotional support after suicidal conversations, but were conflicted around AI-based technologies, including trade-offs between potential benefits of ML-based chatbots for training and limitations of prediction or real-time response suggestions, due to the sensitive, context-dependent decisions that volunteers must make. Our work has important implications for the nuanced considerations and design choices for developing online mental health technologies.

‍
# AI Chatbot, Online Counseling, Human-AI Collaboration