Session termination as a crisis response has a structural vulnerability that none of the consumer products have solved: A user can immediately open a new chat. Across our testing, a user who had disclosed suicidal ideation, completed a safety plan, or triggered an escalation protocol could restart a fresh conversation with no continuity of the prior crisis state—effectively resetting the clinical record.
This is an architectural failure, not a content failure. The stateless session model — where each conversation starts fresh — is fundamentally incompatible with crisis care, which requires longitudinal risk tracking. A safety plan completed in one session that disappears in the next isn't protective; it's theater. The very design pattern that makes chatbots easy to use (no account, no memory required) is what makes them clinically dangerous in crisis contexts.