A Passive, Offline Crisis Detection Framework for Mental Health Safety Using Everyday Devices

This research contributes a passive, offline, privacy-preserving conceptual framework (SAFY) for detecting potential mental health crises using everyday devices. It integrates behavioral and physiological signals with personalized baseline monitoring and user-informed, low-effort interventions, followed by optional escalation to a trusted person. The study also grounds the framework in survey findings, highlighting users’ needs for privacy, emotional authenticity, autonomy, and support without active engagement.