From algorithms to care: A bibliometric overview of AI research in mental health

Artificial intelligence (AI) is increasingly used in mental health research, yet the literature remains dispersed across clinical and computational domains. This study provides a bibliometric analysis of English-language journal articles on AI in mental health indexed in the Web of Science Core Collection between 2021 and 2025. Using Bibliometrix and VOSviewer, we analyzed 1917 articles through descriptive indicators and science-mapping techniques. The results show rapid publication growth across an interdisciplinary literature spanning psychiatry, psychology, neuroscience, and computer science. China and the United States lead in total publications and total citations, with the United States showing extensive international co-authorship connections. The literature is heavily concentrated on depression and on detection or prediction tasks, while the use of deep learning, multimodal data, and transformer-based natural language processing has increased. The mapped literature suggests a field that is expanding at the intersection of clinical and computational research, while revealing patterns relevant to ongoing discussions of clinical validation, unequal global representation, model interpretability, and translation into practice.