Abstract:

Artificial intelligence is increasingly entering palliative care and social work, yet bringing technology into care does not necessarily mean that it becomes part of care. Existing studies have largely followed two lines of inquiry. One highlights the potential of AI to improve efficiency, prediction, and service delivery, while the other focuses on ethical concerns such as surveillance, algorithmic bias, and dehumanization. More recent work in Science and Technology Studies (STS) has moved beyond this opposition by treating AI as a socio-technical actor involved in human-technology relations. Even so, relatively little attention has been paid to how AI becomes embedded in everyday care practice and how human-AI coexistence is gradually produced through interaction, negotiation, and organizational adaptation, particularly in emotionally demanding and ethically sensitive settings such as palliative care.This article examines these questions through an AI-enabled palliative care social work intervention undertaken in a tertiary hospital in China. Rather than evaluating whether AI improves care outcomes, the study focuses on the practical processes through which AI is incorporated into care relationships. Qualitative inquiry was embedded within the intervention evaluation and drew on interviews with patients, family caregivers, social workers, healthcare professionals, and AI developers, together with participant observation, system interaction records, and field notes. This research design made it possible to follow AI as it moved through everyday practice and to examine how its role was continuously shaped by different actors.The findings suggest that AI does not enter care as a stable or predetermined technological actor. Instead, its place within palliative care is progressively established through ongoing adjustments to local contexts, emotional exchanges, and ethical deliberation. As AI is adapted to existing routines, communication styles, and professional practices, it gradually shifts from a digital tool to a participant in care relationships. At the same time, emotional support becomes reorganized rather than automated. AI helps sustain communication, encourages emotional expression, and extends opportunities for interaction, while social workers remain responsible for interpretation, relationship building, and professional judgement. Ethical responsibility also emerges as a negotiated accomplishment. Decisions about risk, intervention, and accountability continue to rely on professional deliberation, with AI functioning as a source of support rather than an autonomous decision-maker.Based on these empirical observations, the article develops a framework of adaptation—senti-modulation—co-governance framework to explain the making of human-AI coexistence in palliative care social work. The framework argues that coexistence should not be understood simply as collaboration between humans and technology. Rather, it is a socially situated process through which AI acquires presence, relational significance, and legitimacy in practice. By shifting attention from technological capability to the social conditions that make coexistence possible, this study contributes to sociological debates on digital care, extends STS discussions of human-AI relations, and offers an empirically grounded account of how AI is domesticated, negotiated, and governed within contemporary care practices.

Key words: human-AI coexistence, artificial intelligence, palliative care, social work, digital care