Pai.ACT: AI-Driven ACT Chatbot for Mental Health Triage and Service Evaluation

Parents of children with special needs in Hong Kong often face limited psychological support, which can negatively impact the child rehabilitation process and the well-being of parent-child relationships. To address this gap, we have developed Pai.ACT, the first deep learning-based mental health advisory system for parents. Pai.ACT features an AI chatbot that integrates the counselling principles of Acceptance and Commitment Therapy (ACT) through natural language processing, providing parents with a human-like voice-to-text experience. Using data from chatbot interactions, the Pai.ACT platform offers assessments regarding the individual's psychological inflexibility status and delivers stratified mental health interventions by: * Low-risk: Users access self-help ACT digital modules tailored to their specific psychological inflexibility processes. * Moderate-risk: In addition to the self-help modules, users receive 4-6 sessions of video-conferencing-based ACT interventions (45-60 minutes per session) conducted by our trained counseling team. * High-risk: Users are directed to specialized mental health services provided by collaborating units. The study includes a regional randomised controlled trial (RCT) in Hong Kong's Sha Tin District, in collaboration with the Shatin District Office. The goal of this regional study is to evaluate the feasibility, acceptability, and potential efficacy of combining AI-driven mental health support across all of Hong Kong. Focus group interviews will also explore parents' perceptions of Pai.ACT and help identify the most effective service model for scaling its use. Pai.ACT provides accessible and comprehensive mental health services to Chinese-speaking parents, helping to alleviate the psychological burden of caregiving. By integrating mental health support with child rehabilitation services and non-governmental organisations, Pai.ACT has the potential to enhance family caregivers' well-being, reduce stigma associated with special