Designing Automation Boundaries for Trustworthy Smart Medication Support

2026-06-27

AI Summary

This study addresses the imbalance between automation levels in intelligent medication systems and older adults’ needs for trust, perceived control, and privacy. It proposes a calibration framework grounded in task risk, user control, and ethical acceptability. Using a mixed-methods approach—combining quantitative experiments with semi-structured interviews—the research evaluates three automation modes: confirmation-required, automatically logged but revocable, and fully automatic. Findings indicate that higher automation does not necessarily enhance trust; participants favored designs that alleviated daily burdens while preserving opportunities for correction. The fully automatic mode performed worst in autonomy, transparency, and user satisfaction. The study further uncovers heterogeneous preferences among older adults regarding privacy sensitivity, digital self-efficacy, and reliance on care, offering empirical evidence and a theoretical foundation for designing age-inclusive intelligent systems.

Abstract

Smart medication systems increasingly automate medication recognition, reminders, and logging. However, automation in home medication routines should be carefully bounded, as users may have different capabilities, privacy expectations, and needs for control over decisions. We present a mixed-methods study of a Smart Medication Support system comparing three automation conditions: confirmation required, automatic logging with undo, and fully automatic support. Across 53 participants and interviews with 11 older adults, we found that higher automation did not necessarily lead to higher trust or acceptance. Participants preferred automation that reduced routine effort while preserving opportunities for correction. Fully automatic support was less interruptive but was rated lower in autonomy, trust, transparency, dignity, and satisfaction. Interviews also showed clear differences among older adults. Their preferences were shaped by privacy concerns, digital confidence, perceived vulnerability, and caregiver involvement. We contribute empirical evidence and design implications for calibrating automation in smart medication systems according to task risk, user control, and ethical acceptability.

Problem

• automation boundaries
• smart medication systems
• trust
• user control
• older adults

Innovation

• automation boundaries
• trust calibration
• smart medication systems
• user control
• ethical acceptability