Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 28, 2026

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
07:14

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models

Published on: December 23, 2025

Safety Mechanisms and Risk Mitigation in Generative AI Mental Health Chatbots: A Systematic Scoping Review.

Lotenna Olisaeloka1,2, Chris G Richardson1,2, Angel Y Wang1

  • 1Department of Psychiatry, Faculty of Medicine, University of British Columbia, Vancouver, BC V6T 0A6, Canada.

Healthcare (Basel, Switzerland)
|May 27, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Psychosurgery01:30

Psychosurgery

Psychosurgery, the surgical alteration or permanent removal of brain tissue to alleviate severe psychological conditions, stands as one of the most radical and controversial treatments in the history of mental health care. Its development and application have evolved significantly, marked by dramatic shifts in scientific understanding and ethical perspectives.
Historical Development of Psychosurgery
In the 1930s, Portuguese neurologist Antonio Egas Moniz introduced a surgical procedure designed...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Adherence to Lower-Risk Cannabis Use Guidelines among Canadian college students: a regression analysis.

Health promotion and chronic disease prevention in Canada : research, policy and practice·2026
Same author

Artificial intelligence (AI) psychosis: mechanisms, clinical risks and safety considerations in generative AI chatbots.

BJPsych open·2026
Same author

Patterns and predictors of the transition between minimally adequate treatment and effective treatment coverage for mental disorders: results from the World Mental Health Survey.

International journal of mental health systems·2026
Same author

Gender, economic stressors, and altered food intake during the COVID-19 pandemic: findings from four waves of a nationally representative cross-sectional survey.

Journal of public health (Oxford, England)·2026
Same author

User experience and safety of generative AI-based mental health chatbots: Scoping review protocol.

PloS one·2026
Same author

Anxiolytic Medication Use in Low- Middle- and High-Income Countries: A World Mental Health Surveys Report.

Human psychopharmacology·2025
Same journal

Correction: Haddock et al. <i>Imagine the Possibilities Pain Coalition</i> and Opioid Marketing to Veterans: Lessons for Military and Veterans Healthcare. <i>Healthcare</i> 2025, <i>13</i>, 434.

Healthcare (Basel, Switzerland)·2026
Same journal

Macro Responsibility in the Microvascular World: Nurse Experiences in Flap Care, a Phenomenological Study.

Healthcare (Basel, Switzerland)·2026
Same journal

Agreement Between Standing Eight-Point Multifrequency Bioelectrical Impedance Analysis and Dual-Energy X-Ray Absorptiometry for Body Composition Assessment in Apparently Healthy Greek Adults.

Healthcare (Basel, Switzerland)·2026
Same journal

'It's Not About the Food'-Understanding the Lived Experience of Patients Who Developed Hospital-Acquired Malnutrition (HAM) and That of Their Carers.

Healthcare (Basel, Switzerland)·2026
Same journal

Unveiling the Humanizing and Therapeutic Values of Live Music in Healthcare Settings: A Scoping Review.

Healthcare (Basel, Switzerland)·2026
Same journal

Respiratory Rehabilitation and Decannulation in Adults with Prolonged Mechanical Ventilation After Tracheostomy: A Narrative Review.

Healthcare (Basel, Switzerland)·2026
See all related articles

Generative AI (GenAI) mental health chatbots need strong safety measures. This review found technical safeguards and user input are crucial, but human oversight and crisis protocols need improvement for safer mental healthcare delivery.

Area of Science:

  • Digital Health
  • Artificial Intelligence in Healthcare
  • Mental Healthcare Technology

Background:

  • Generative AI (GenAI) chatbots offer potential for mental healthcare but pose risks due to open-ended outputs.
  • Existing safety implementations in purpose-built GenAI mental health interventions are not well understood.
  • General-purpose GenAI chatbot harms necessitate dedicated safeguards for specialized applications.

Purpose of the Study:

  • To conduct a scoping review of safety mechanisms and risk mitigation strategies in GenAI mental health chatbots.
  • To synthesize findings on technical safeguards, pre-deployment considerations, and delivery-phase risk management.
  • To identify gaps and inform future development and regulation of these interventions.

Main Methods:

  • Scoping review following Joanna Briggs Institute and PRISMA-ScR guidelines.
Keywords:
AI chatbotsAI safetyChatGPTartificial intelligencegenerative AImental health chatbots

Related Experiment Videos

Last Updated: May 28, 2026

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
07:14

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models

Published on: December 23, 2025

  • Systematic search of seven academic databases and search engines (July 2025).
  • Narrative synthesis of safety mechanisms across technical, pre-deployment, and delivery domains.

    • Main Results:

    • Twenty-one studies from 11 countries were included.
    • Fine-tuning and prompt engineering were common technical safeguards; layered architectures were less frequent.
    • Pre-deployment included co-design and data privacy, but human oversight and crisis protocols were limited, with documented failures like missed suicidal ideation.

    Conclusions:

    • GenAI mental health chatbots require a sociotechnical approach integrating technical safeguards, user co-design, procedural controls, and human oversight.
    • Further research is needed to evaluate efficacy, enhance safeguards, and standardize safety outcome measurement.
    • Regulatory oversight is essential for safe integration into mental healthcare systems.