Seeking Late Night Life Lines: Experiences of Conversational AI Use in Mental Health Crisis

TL;DR

This work investigates how people experience conversational AI during mental health crises and defines responsible AI behavior through a bridging lens. Using a mixed-methods design with $n=53$ lived-experience testimonials and $n=16$ expert interviews, the authors apply the stages-of-change framework to interpret how AI can move users from contemplation toward action while de-escalating harmful intents. Key contributions include a narrative portrait of AI crisis use, a staged interpretation of AI’s role as a bridge to human support, and three design guidelines for preparedness-building, leveraging machine strengths, and safeguarding against over-reliance. The study highlights the potential for AI to facilitate immediate relief and prepare users for subsequent human care, while underscoring governance, equity, and safety considerations to prevent misuses and replace rather than supplement traditional crisis resources.

Abstract

Online, people often recount their experiences turning to conversational AI agents (e.g., ChatGPT, Claude, Copilot) for mental health support -- going so far as to replace their therapists. These anecdotes suggest that AI agents have great potential to offer accessible mental health support. However, it's unclear how to meet this potential in extreme mental health crisis use cases. In this work, we explore the first-person experience of turning to a conversational AI agent in a mental health crisis. From a testimonial survey (n = 53) of lived experiences, we find that people use AI agents to fill the in-between spaces of human support; they turn to AI due to lack of access to mental health professionals or fears of burdening others. At the same time, our interviews with mental health experts (n = 16) suggest that human-human connection is an essential positive action when managing a mental health crisis. Using the stages of change model, our results suggest that a responsible AI crisis intervention is one that increases the user's preparedness to take a positive action while de-escalating any intended negative action. We discuss the implications of designing conversational AI agents as bridges towards human-human connection rather than ends in themselves.

Figures (3)

• Figure 1: Overview or our narrative analysis method. On the left is an illustrative example of a survey response. On the right is the narrative analysis method consisting of four components and themes we identified from our responses.
• Figure 2: Results of self-reported availability of mental health resources and likelihood the respondent would use given resource. While we find the obvious result of AI tools being the most available and likely-to-use resource, our respondents report significant availability of human-based resources, such as family and friends.
• Figure 3: Stages of change model applied to our findings. Our interview and survey results synthesize into a recommendation of AI chatbots serving as a bridge towards positive action, such as human-human connection.