8 artifacts matching
Framework
Preparing AI chatbots to respond to patient distress and suicidality in high-risk healthcare settings
Comment describing the suicide-risk and distress safety architecture built for 'Suzy', a generative AI chatbot offering recovery, wellness and local-resource support to adults receiving medication tr…
Peer-reviewed
Safety, efficacy and acceptability of human-GenAI single-session exposure-based intervention for academic anxiety: randomized controlled trials
Two preregistered randomized controlled trials (N = 425 in total) of a hybrid human-generative-AI single-session exposure program for academic anxiety, evaluated against waitlist and active controls.…
Peer-reviewed
Exploring generalizability and explainability of LLMs in classifying clinically rated suicidal ideation using heterogeneous data
Hong Kong study asking whether a language-model classifier of clinician-rated suicidal ideation performs unequally across patient subgroups because of linguistic heterogeneity. Cantonese clinical-int…
Peer-reviewed
A scoping review on the mental health harms of LLM-based chatbots
A PRISMA-based scoping review synthesising research on mental health harms associated with chatbots built on large language models. A systematic search with a validated search string across five data…
Peer-reviewed
Real-world use of large language models for mental health in 2024
Survey of 1,871 US adults conducted between August and October 2024, using stratified sampling across age, sex and race/ethnicity to approximate national demographics, measuring how many people use g…
Peer-reviewed
AI chatbots and youth suicide risk: current evidence, critical gaps, and a clinical research agenda
Review by researchers at Crisis Text Line examining how general-purpose chatbots and AI companions detect and respond to suicide-risk disclosures from young people, and what the existing evidence can…
Peer-reviewed
Safety boundary maintenance in consumer AI systems responding to pediatric health queries: a cross-platform benchmark evaluation under naturalistic and adversarially pressured conditions
Benchmark evaluation of how four consumer AI systems maintain safety boundaries when answering paediatric health questions from caregivers, using PediatricSafetyBench-v2: 300 authentic caregiver quer…
Peer-reviewed
An AI-based mental health guardrail and dataset for identifying psychiatric crises in text-based conversations
Peer-reviewed evaluation of the Verily Mental Health Guardrail (VMHG), an AI-based classifier for identifying psychiatric crises in text-based conversations with language models. The guardrail was ev…