ScholarWorks@성균관대학교

초록

Online mental health assessment systems offer promise for individuals to evaluate their mental health without social stigma. With recent advancements, these systems evolved beyond pre-defined questionnaires to detect mental health conditions from user-generated text. However, existing research focused on model accuracy, with limited attention to user experiences. To bridge these gaps, we examine users’ intention to adopt AI-based mental health assessment systems and investigate how symptom-based approaches affect user experience. We developed a mental health assessment system using natural language processing and conducted a within-subject study with 30 participants. Results demonstrated that symptom-based explanations enhance user’s understanding of their mental health, with most participants expressing their intention to use. While accessibility, anonymity, and self-reflection positively influenced usage intention, the generalized result and lack of detailed explanation were a limiting factor. The findings suggest AI-based mental health assessment systems as supportive tools for early-stage evaluations, emphasizing the importance of personalized assessment. © 2025 Copyright held by the owner/author(s).

키워드