Designing Reflective Learning Evidence for Older Adults: An Ethically Constrained AI-Supported Environment

Abstract

As societies continue to age, sustaining reflective learning and health-related self-management among older adults has become a critical issue in lifelong learning. This study reports a design-oriented, practice-based investigation of a reflective learning environment developed alongside a community-based frailty prevention program. Rather than evaluating intervention effectiveness, the study examines how Learning Analytics can function as a design lens for structuring learner-generated reflective evidence. A simple digital reflection record system was implemented to accumulate weekly reflections on daily practices and perceived changes. Learning Analytics is treated not as post-hoc analysis but as the prior design of analyzable reflective traces. Artificial intelligence functions are deliberately limited to background organization of records without algorithmic interpretation or automated feedback. This limitation represents a theoretical position that rejects algorithmic interpretation in favor of human meaning-making, rather than a mere precaution for ethical sensitivity. The contribution lies in stabilizing reflective traces while preserving learner autonomy and ethical restraint in later-life learning contexts.