Related Experiment Videos
Exploring Applications of AI in the Crisis Line Sector: Protocol for a Scoping Review
Allison Crawford1,2, Mackenzie Earle1, Gisell Castillo1
1Centre for Addiction and Mental Health, Toronto, ON, Canada.
Background:
Crisis helplines are a vital component of a robust public health approach to suicide prevention as they are often free, accessible, and provide immediate support to individuals in distress. AI presents an opportunity for novel applications to support and improve crisis line services across a variety of functions, including assessing suicide risk, identifying issues, tracking responder behaviors, and providing prompts and reminders. However, the use of AI in the crisis sector also raises critical questions regarding safety, ethics, privacy, efficacy, feasibility, and acceptability among interest holders. The extent to which AI is currently being explored and implemented in crisis line contexts is unknown.
Objective:
The objective of this scoping review is to examine how AI is being used in crisis line services and to explore the perspectives of different interest holders regarding its application.
Methods:
This scoping review will be conducted in accordance with a validated methodological framework and reported using the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews). A comprehensive search strategy will be developed with an experienced health sciences librarian and implemented in multiple databases, including Ovid MEDLINE, Embase, APA PsycInfo, CINAHL, IEEE Xplore, ACM Digital Library, and Web of Science. We will include primary research describing AI applications and implementation in the crisis line sector and studies exploring interest holder perspectives. Screening of titles and abstracts and of full texts will be performed independently and in duplicate following a screening calibration process. Disagreements will be resolved through discussion. Expert opinions will be sought as needed to make final determinations regarding included and excluded sources. A data extraction form will be created and piloted to capture information on study characteristics, study findings and outcomes, AI use cases, and considerations for various concepts relevant to crisis line services (eg, safety, trust, equity, ethics, and privacy). Extracted data will be synthesized using descriptive and narrative approaches.
Results:
As of June 2026, the search strategy has been finalized and translated across databases. Screening of 3380 unique records is ongoing and will conclude by September 2026. Results will be available by winter 2027.
Conclusions:
AI is changing public mental health by creating new opportunities to optimize interventions and enhance care. However, several implementation challenges remain regarding the use of AI applications in the crisis sector. This review is a timely and comprehensive exploration into the state of AI in the crisis sector and will identify existing knowledge and evidence gaps to help inform practice, policy, and future research. By systematically mapping current AI applications; interest holder perspectives; and how core domains such as safety, privacy, and equity are addressed, this review will provide a foundational evidence base to guide responsible, person-centered AI integration and a robust research agenda tailored to crisis line contexts.
International Registered Report Identifier (Irrid):
PRR1-10.2196/95087.
