Protective Interruption

Description

AI or human intervention redirects the person toward support, delay, help-seeking, or safety planning.

Rationale

Protective interruption — whether by AI safety features, an alert system, a concerned third party, or the person's own ambivalence — is a critical outcome variable. Documenting when and how interruptions occurred, whether they were effective, and what the person's response was is necessary for understanding the current risk state and designing future interventions.

Evidence Base

• Inside a Mass Shooter’s Harrowing History With ChatGPT

  Mark Follman, 2024. Inside a Mass Shooter’s Harrowing History With ChatGPT. Mother Jones.

  News / journalism

• AI-associated delusions: the missing dimension of mortality

  Idris Altinisik, Utku Igcioglu, Murat Emul, 2024. AI-associated delusions: the missing dimension of mortality. Lancet Psychiatry, Vol 13, pages 456. https://doi.org/10.1016/S2215-0366(25)00396-7.

  Commentary / Editorial

• Responses of AI chatbots to escalating suicide risk: A simulation study of repeated interactions

  Wojciech Pichowicz, Marek Kotas, Natalia Kalka, Julia Dembowska, Julia Karska, Patryk Piotrowski, Bła˙zej Misiak, 2026. Responses of AI chatbots to escalating suicide risk: A simulation study of repeated interactions. Journal of Affective Disorders, 416, 122586. https://doi.org/10.1016/j.jad.2026.122586

  Peer-reviewed study (original research)

• Human-in-the-Loop Suicide Risk Detection and Clinical Escalation in a Generative AI Mental Health Chatbot: Prospective Implementation Study

  Suzette Glasner, Darcy Michero, Matthew Stewart, 2026. Human-in-the-Loop Suicide Risk Detection and Clinical Escalation in a Generative AI Mental Health Chatbot: Prospective Implementation Study. JMIR Preprints. DOI: https://doi.org/10.2196/preprints.106118.

  Preprint