Review of Artificial Intelligence Applications in Mental Health Diagnosis and Therapy

Abstract

Mental illness represents a growing global health concern, with increasing prevalence and burden across populations. Timely diagnosis and effective treatment are essential to improving individual health outcomes. In this context, artificial intelligence (AI) offers significant potential through rapid and precise data analysis. This narrative literature review aims to examine the application of AI technologies, specifically natural language processing (NLP), machine learning (ML), and predictive modelling, in the diagnosis and therapy of mental health disorders among adult clinical populations. The review includes twelve peer-reviewed articles indexed in Scopus between 2014 and 2024, selected based on relevance, methodological rigor, and contribution to the field. Findings were organised into three thematic clusters: (1) AI-assisted diagnostic accuracy and predictive modelling (e.g., fMRI-based PTSD prediction, depression detection via multimodal neural networks); (2) AI-enhanced therapeutic delivery and user engagement (e.g., AI-assisted online social therapy, remotely supervised brain stimulation); and (3) ethical, privacy, and implementation challenges (e.g., data bias, lack of transparency, and population representativeness). These technologies show promise in reducing human error and enhancing mental health care delivery; however, persistent challenges include data privacy, ethical considerations, and the need for diverse, large-scale datasets. Future research should focus on developing standardised implementation protocols using frameworks such as PRISMA, ensuring population diversity, and addressing ethical safeguards. Collaboration among mental health professionals, AI technologists, and policymakers is essential to promote safe, effective, and equitable integration of AI in psychological diagnosis and therapy.