Year: 2026
Type: dissertation
Abstract: This dissertation proposes a practical method for reflection support in elderly care settings, positioning AI not as a substitute for human judgment or meaning making, but as a "reflective mediator" that visualizes care workers' reflections, with managers using these visualizations as resources for dialogue.While AI use in the care sector has primarily focused on operational efficiency and decision support, less attention has been paid to how organizationally shared purposes are connected to individual learning and everyday practice.This study addresses this issue by focusing on AI-mediated visualization of reflection and the dialogical role of managers, presenting an operational configuration that can be repeatedly implemented in practice.The research was conducted in a day-care facility for older adults, where AI was used to organize and visualize care workers' daily reflections, and managers employed these visualizations as resources for dialogue.Changes in care workers' patterns of action, their understanding of work purposes, and teamwork were documented using a combination of quantitative indicators and qualitative analysis of reflective descriptions.Through the operation, reflective descriptions initially centered on tasks and events gradually shifted to include the purposes of actions through continued reflection and dialogue with managers, and this shift was maintained over time.These findings suggest that a viable operational configuration can be established in which AI supports behavioral adjustment not by directing human judgment, but by enabling care workers to reframe their actions within an organizational context through manager-mediated shared reflection.
Author: 三ツ木, 直樹
Institution: Japan Advanced Institute of Science and Technology
Language: Japanese
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7164797449
