CHATGPT-BASED PSYCHOLOGICAL SUPPORT AND MENTAL HEALTH OUTCOMES AMONG ARAB ADULTS: THE MEDIATING ROLES OF EMOTIONAL SELF-DISCLOSURE, PERCEIVED USEFULNESS, AND SELF-EFFICACY

Abstract

This study examines the association between ChatGPT-based psychological support and depression, anxiety, and stress among Arab adults, and investigates the potential mediating roles of emotional self-disclosure, perceived usefulness, and general self-efficacy. It adopts a mixed-methods, cross-sectional design. The quantitative component includes 512 Palestinian Arab adults living in the study area described in this thesis as Palestine 1948, all of whom report prior use of ChatGPT for emotional or psychological support. The study measures perceived ChatGPT-based psychological support, distress disclosure, perceived usefulness, general self-efficacy, and symptoms of depression, anxiety, and stress. Quantitative analyses test direct associations and parallel and sequential indirect pathways. The qualitative component comprises semi-structured interviews with 16 additional adults from the same target population. These interviewees are distinct from the 512 survey respondents; therefore, 528 individuals participate across the two study components. Thematic analysis explores interviewees’ experiences and helps contextualize the quantitative findings. Results indicate that stronger perceived ChatGPT-based psychological support is associated with lower depression, anxiety, and stress symptoms, with the strongest association observed for stress. The full sequential mediation pathway is not supported. In the parallel model, general self-efficacy is the only significant mediator across all three outcomes; emotional self-disclosure and perceived usefulness do not show significant independent mediation. The qualitative findings highlight immediate access, non-judgmental interaction, emotional validation, and practical guidance, alongside concerns about privacy, trust, cultural relevance, and the absence of human warmth. These findings suggest that perceived psychological support through ChatGPT may be associated with mental health outcomes partly through users’ beliefs in their ability to manage difficulties. Because the design is cross-sectional, the findings do not establish causality or demonstrate that ChatGPT is a safe or clinically effective treatment. The study supports cautious exploration of AI as an informal adjunct, not a substitute for professional mental health care.