AI-Assisted Human-Computer Interaction for Loneliness and Social Isolation Intervention in Older Adults: A Scoping Review

With the accelerating global aging, loneliness and social isolation among older adults have become prominent public health issues, linked to cognitive decline, depression and higher mortality. Rooted in factors like bereavement, physical impairment and digital divide, they are hardly fully addressed by traditional interventions alone. The rapid development of AI and HCI technologies has brought new solutions. AI-assisted HCI models (e.g., voice interaction, virtual companions, social robots) provide personalized, low-threshold social support and emotional companionship, becoming a research hotspot in gerontology, nursing and public health. However, current research remains fragmented, with diverse AI types, HCI modes and intervention scenarios, leading to inconsistent conclusions. A lack of systematic collation hinders researchers and policymakers from grasping the research status and gaps, limiting technology popularization. To achieve the above research purposes, this scoping review will focus on addressing the following specific research questions, which are designed to systematically clarify the current research status and fill existing gaps: First, what are the main types, core technical characteristics and interaction modes of AI-assisted HCI models applied in the intervention of loneliness and social isolation among older adults? Specifically, it will explore the classification of related models (such as voice interaction, virtual companion robots, intelligent remote social platforms, etc.) and their applicable technical paths. Second, what are the application scenarios (such as home care, community services, institutional care) and target populations (such as healthy older adults, frail older adults, older adults with chronic diseases) of these AI-assisted HCI intervention models, and what are the differences in their application effects among different groups and scenarios? Third, what are the key influencing factors of the intervention effects of AI-assisted HCI models on loneliness and social isolation among older adults, including technical adaptability, user participation degree, intervention duration and other aspects? Fourth, what are the main problems and challenges existing in the current research and application process, such as technical bottlenecks, ethical and privacy risks, digital divide constraints, and insufficient research depth? Fifth, based on the existing research gaps, what are the feasible future research directions and practical optimization suggestions for AI-assisted HCI intervention models in this field? Guided by these research questions, the specific content of this study will include clarifying core concepts, systematically retrieving and sorting out relevant literature, classifying and evaluating intervention models, analyzing existing challenges, and putting forward targeted suggestions. This review, focusing on the intersection of AI, HCI and gerontological public health, fills the gap of systematic summary, provides practical guidance for technology optimization, and contributes to healthy aging.