ARTIFICIAL INTELLIGENCE–ENHANCED MINDFULNESS-BASED COGNITIVE BEHAVIOR THERAPY (AI–MB-CBT) FOR ADULT MENTAL HEALTH EFFECTIVENESS, MECHANISMS, AND PERSONALIZATION

Abstract

This study evaluated the effectiveness, mechanisms, and personalization effects of Artificial Intelligence–Enhanced Mindfulness-Based Cognitive Behavior Therapy (AI–MB-CBT) for adult mental health using a quantitative longitudinal design. A total of 312 adults participated (mean age = 34.8 years, SD = 9.6; 56.1% women), drawn from community (62.8%) and outpatient (37.2%) contexts. Outcomes and process variables were assessed at baseline, mid-treatment, post-treatment, and follow-up, alongside platform-logged engagement and AI adaptation indicators. Descriptive results showed clinically meaningful symptom reductions over time. Depression severity decreased from baseline (M = 15.9, SD = 6.2) to post-treatment (M = 10.4, SD = 5.8) and remained lower at follow-up (M = 10.8, SD = 6.0). Anxiety severity declined from baseline (M = 14.7, SD = 5.8) to post-treatment (M = 9.8, SD = 5.5) and follow-up (M = 10.1, SD = 5.7). Stress/distress improved from baseline (M = 19.8, SD = 6.4) to post-treatment (M = 14.2, SD = 6.1), with follow-up stability (M = 14.6, SD = 6.3). Functional impairment improved from baseline (M = 17.1, SD = 6.9) to post-treatment (M = 13.4, SD = 6.7) and follow-up (M = 13.1, SD = 6.6). Mechanisms shifted in theoretically consistent directions, including increases in mindfulness skills (baseline M = 3.1, SD = 0.7; post-treatment M = 3.7, SD = 0.7) and decreases in rumination/worry (baseline M = 3.6, SD = 0.8; post-treatment M = 3.0, SD = 0.8). Engagement was substantial (module completion M = 61.4%, SD = 23.6), and adaptation exposure varied (adaptations M = 5.2, SD = 3.4; time-to-first adaptation M = 9.1 days, SD = 6.8). Overall findings supported AI–MB-CBT as an effective, mechanism-linked, and personalization-sensitive intervention for adult mental health.