AI for Relational Reconnection and Social Support in Alzheimer’s Disease: A Conceptual Framework for Socially Embedded Systems

Alzheimer's Disease (AD) has traditionally been approached through a biomedical lens, focusing on neurodegenerative markers such as amyloid-β plaques and tau protein accumulation. However, clinical evidence increasingly demonstrates that social dysfunction, which including identity confusion, emotional withdrawal, and breakdowns in social roles. This article reconceptualizes Alzheimer's disease as a disorder in which the primary dimension of decline lies in social relationship management capacity (SRMC), while recognizing that neurobiological and cognitive deterioration remain integral to its manifestation and progression. SRMC refers to a person's ability to identify, interpret, maintain, and regulate social ties embedded in complex networks. This article introduces a conceptual and technical framework for a socially embedded artificial intelligence (AI) framework designed to recognize and compensate for the deterioration of SRMC in AD. Drawing on social capital theory, affective computing, and neural social cognition research, this framework proposes a four-dimensional intervention model: relationship recognition, relationship learning, relationship establishment, and relationship management. By aligning cutting-edge AI techniques with the lived social reality of individuals with AD, this approach not only provides a new path for supportive care but also reorients ethical and technological discourse toward sustaining social personhood in the face of neurodegeneration.