AI chatbots vary in how they respond when suicide help is declined
What do AI chatbots do after users decline therapeutic redirection for suicide thoughts?
Human-Computer InteractionArtificial IntelligenceComputers and Society
Summary
AI chatbots usually suggest that people with suicidal thoughts seek help from crisis lines or humans. This study found that when users say no to these suggestions, chatbots respond very differently: some keep encouraging help many times, while newer chatbot versions tend to give up more quickly. The chatbots also often try to support the user themselves instead of pushing for outside help. However, very few chatbots ask why the user refuses help. The authors suggest safety tests should check chatbot behavior after help is refused, not just whether help is offered.
What this means in practice
- •For mental health app developers: Design chatbot behaviors that better handle refusals of crisis referrals by tracking persistence and user support strategies.
- •For customer support chatbot teams: Improve chatbot crisis management by monitoring responses after users reject recommended referrals, ensuring safer engagement strategies.
Authors
Subigya Nepal, Ruth Heo, Emma M Parrish, Colin A Depp
Abstract
Safety evaluations of artificial intelligence (AI) chatbots focus on therapeutic redirection, statements urging users who disclose suicide thoughts to seek help from crisis lines or other human sources. How chatbots respond when users refuse that help is unclear. This cross-sectional audit ran 2,550 scripted multi-turn conversations in September 2026. Each conversation opened with a disclosure of suicide thoughts at 3 levels of severity, answered every referral with up to 10 standardized refusals and then changed the subject. All 2,550 conversations included a referral, 2,542 (99.7%) in the first reply. However, mean persistence ranged from 0.01 to 9.94 refusals across chatbots. Newer versions of all 5 chatbots tested in 2 versions persisted less than older versions (each P < .05). Compared with passive ideation, a disclosure of intent was associated with 1.26 more refusals sustained (95% CI, 1.06-1.46) and higher odds of renewed referral after the topic change (odds ratio, 10.07; 95% CI, 7.43-13.66). Chatbots that persisted less in referral more often presented themselves as the user's support (Spearman rho = -0.80; P < .001). Of 8,818 refusal-phase replies urging outside help, only 173 (2.0%) asked about barriers to seeking it. Chatbots uniformly recommended outside help after disclosures of suicide thoughts but varied widely in persisting in those recommendations, with newer versions persisting less. Safety evaluations should test what happens after a crisis referral is refused, not only whether one is offered.