Suicide-Related Responses From AI Chatbots Through Application Programming Interfaces and Consumer-Facing Apps: Observational Cross-Sectional Comparative Study - Report - MDSpire
Clinical Report: AI Chatbot Responses to Suicide-Related Prompts
Overview
The article frames suicide-related chatbot responses as a safety concern amid growing use of generative AI for health and emotional support. It notes that prior research has found variable responses to suicide-risk and self-harm prompts, ranging from crisis-resource referrals to potentially inappropriate or harmful content; the supplied material does not include the comparative study’s methods or results.
Background
Suicide remains a substantial public health concern in the United States, with approximately 49,000 deaths reported for 2023 in the article’s cited CDC source. Limited access to in-person mental health care has contributed to reliance on telehealth, crisis lines, and other technology-enabled support. Generative AI chatbots are increasingly used for health-related and emotionally salient questions, and their availability and conversational format may make them an initial source of support for people experiencing suicidal thoughts or behaviors. WHO safe-messaging guidance advises against detailed descriptions of suicide methods and emphasizes referral to crisis resources.
Data Highlights
The supplied article context does not provide the comparative study’s sample, chatbot platforms, measures, numerical results, or statistical analyses. It reports approximately 49,000 U.S. suicide deaths in 2023 and cites an estimate that health topics account for approximately 5% of ChatGPT messages.
Key Findings
The article identifies suicide as a major U.S. public health concern and cites approximately 49,000 deaths in 2023.
The introduction describes increasing use of generative AI chatbots for health-related inquiries and emotional or psychological guidance.
The article cites research reporting that chatbots can receive higher empathy ratings than human clinicians, while also noting concerns about sycophantic responses in suicide-related exchanges.
Prior research cited in the article found that LLM responses to suicide-risk, self-harm, and mental health crisis prompts vary.
Some previously studied chatbot responses appropriately direct users to crisis resources, while others provide content characterized in the article as potentially clinically inappropriate or harmful.
WHO safe-messaging guidance advises against detailed accounts of suicide methods and stresses referral to crisis resources.
Clinical Implications
The supplied material identifies consistency and safety of chatbot responses to suicide-related prompts as concerns, but does not provide findings from the comparative analysis that would support platform-specific clinical conclusions. The cited WHO guidance advises avoiding detailed descriptions of suicide methods and referring people to crisis resources.
Conclusion
The article’s context describes growing use of AI chatbots for emotionally salient and health-related exchanges alongside documented variability in responses to suicide-related prompts. The comparative study’s results are not included in the supplied material.
by Jonathan Cantor, Ellice Huang, Li Ang Zhang, Adrian Salas, Fang Zhang, Alyssa Burnett, Andrew Lohn, Pat Pataranutaporn, Benjamin Rader, Hao Yu, Ryan K McBain
Lower self-esteem and greater romantic loneliness may help explain how perceived phone disengagement relates to psychological well-being in young adults.
