Vulnerability Intersections: An Integrative Conceptual Framework for Understanding Vulnerability in AI-Assisted Care for Older Adults.

Artificial Intelligence (AI) applications are increasingly used to support independent living and care for older adults, particularly those in vulnerable situations. Yet, how vulnerability is constructed when AI enters later-life care remains poorly theorized. Existing approaches to vulnerability in old age inadequately address AI-related risks, structural conditions, exposure pathways, and life course disadvantages. Using a theory-synthesis approach, this article proposes "vulnerability intersections" as an integrative framework for how vulnerability is constructed, enacted, and negotiated in AI-assisted care for older adults. The framework extends existing sociotechnical accounts by conceptualizing vulnerability as taking four overlapping forms: anticipatory, opaque, routinized, and normative. It locates vulnerability factors at the micro level of older adults' lives, the meso level of care organizations and technology providers, and the macro level of political-economic and regulatory arrangements, with life-course accumulation spanning all three. It also proposes three exposure pathways through which vulnerability factors become consequential: AI-mediated information and representations of aging; AI-powered care organization and delivery; and AI-configured power, participation, and social relations. The framework shifts attention away from viewing older adults as inherently vulnerable toward examining the sociotechnical arrangements under which vulnerability is anticipated, classified, distributed, and negotiated when AI enters care.