Personalized for Whom? Examining Individual Differences in AI-Based Mental Health Interventions for Older Adults: A Scoping Review

Artificial intelligence (AI) is increasingly promoted as a scalable way to support older adults’ mental health, yet older adults vary widely in cognitive functioning, digital literacy, and mental health needs. This scoping review examined what types of AI-based mental health interventions are used with older adults and whether they are adapted to individual differences across subgroups. Following PRISMA-ScR guidelines, we searched PsycINFO, PubMed, Scopus, and Web of Science, along with a supplementary Google Scholar search, for English-language empirical studies published from 2023 onward that evaluated an AI-based mental health intervention among adults aged 65 and older, or samples with a mean age of 65 or above. Of 3466 records screened, only two studies met the inclusion criteria: one examining an AI-powered voice assistant and one examining a generative AI chatbot. Findings suggested preliminary benefits, although statistical significance was mixed across the two studies. However, neither study used age subgroup, cognitive functioning, digital literacy, or baseline mental health need to guide intervention design, delivery, or content adaptation, and neither reported subgroup or moderation analyses. The evidence identified by this search is therefore insufficient to support broad claims about AI suitability for older adults’ mental health and does not yet demonstrate meaningful personalization. Future research must clarify for whom AI-based mental health interventions work, under what conditions, and with what adaptations.