Exploring the Role of Artificial Intelligence Applications in Mental Healthcare: A Systematic Review of Patients' and Clinicians' Experiences

Background Artificial intelligence (AI) applications are increasingly being adopted in mental healthcare to improve accessibility, reduce stigma, and enhance treatment efficiency. However, the experiences and perceptions of patients and clinicians regarding these tools remain unclear, with concerns about trust, empathy, privacy, and clinical validity frequently raised in the literature. Aim This systematic review aimed to synthesize existing evidence on patients’ and clinicians’ experiences of AI applications in mental healthcare and to evaluate how these tools impact care delivery, quality, and therapeutic relationships. Methods A systematic review was conducted following PRISMA guidelines. Databases including PubMed, Cochrane Library, and Google Scholar were searched for primary research studies published in English between 2018 and 2025. Eligible studies employed qualitative, quantitative, or mixed-methods designs and examined patient or clinician experiences with AI tools in mental healthcare. Methodological quality was appraised using the Mixed Methods Appraisal Tool (MMAT). Narrative synthesis and thematic analysis were performed to integrate findings across diverse study types. Results From 5,400 initial records, 20 studies met the inclusion criteria. Thematic analysis identified four major themes: (1) AI app engagement and mental health improvement with studies demonstrating symptom reductions and positive user experiences; (2) relational concerns and professional scepticism with highlighting issues of trust, empathy, and privacy; (3) youth attitudes toward AI, where studies found selective willingness to share non-sensitive data; and (4) professional readiness and systemic barriers along with lack of training, regulatory frameworks, and funding. While evidence indicates AI tools can improve accessibility, literacy, and administrative efficiency, persistent concerns limit adoption in direct therapeutic contexts. Conclusion AI applications demonstrate promise as adjuncts in mental healthcare by enhancing accessibility, reducing stigma, and supporting self-management. However, their effectiveness and acceptability are moderated by issues of trust, privacy, cultural sensitivity, and clinician readiness. To achieve sustainable integration, future efforts must focus on developing standardized governance frameworks, culturally inclusive designs, and comprehensive clinician training.