Artificial Intelligence Interventions for Resilience and Mental Health in Older Adults: A Systematic Review

Background: As global population ages, the supply-demand imbalance in geriatric mental health services has become increasingly pronounced. In this emerging field, artificial intelligence (AI)-driven interventions for geriatric mental health remain fragmented. This systematic review synthesizes the application process of AI interventions and dissects the disciplinary knowledge structure of existing studies to unlock AI potential and facilitate evidence-based interventions. Methods: Guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, a systematic search of seven databases was performed from January 1, 2005 to November 5, 2025. This review includes a total of 95 studies in English and Chinese on AI-based interventions for geriatric mental health. These studies were categorized by basic characteristics, intervention processes, and thematic clustering. Results: The field has evolved from an initial exploratory phase to a period of rapid growth. These AI interventions encompass four core dimensions, involving seven AI technology types, nine intervention pathways, and seven key acceptance-influencing factors. Research topics focused on clinical efficacy and risk management of AI interventions, as well as psychological assessment and human-computer interaction. This study summarized interdisciplinary, comprehensive research on AI-based geriatric mental health interventions. Furthermore, it establishes an explanatory model emphasizing multidisciplinary synergy, which provides a framework for future theoretical and practical development. Conclusions: There remains a notable gap between technology design and the psychological characteristics of older adults in AI mental health interventions, and interdisciplinary research is still insufficient. Future efforts should strengthen collaboration across clinical medicine, psychology, sociology, and computer science, embed elderly psychological traits into closed-loop technology design, increase policy support, enhance AI accessibility in community and family settings, and improve caregivers’ technical literacy. Only through these steps can human-centered AI technology sustainably and equitably address the unmet mental health needs of older adults in the context of global aging. Registration: The study has been registered on https://www.crd.york.ac.uk/PROSPERO/ (registration number: CRD420261287044; registration link: https://www.crd.york.ac.uk/PROSPERO/view/CRD420261287044).