Examining and Promoting Explainable Recommendations for Personal Sensing Technology Acceptance

Abstract

Personal sensing is a promising approach for enabling the delivery of timely and personalised recommendations to improve mental health and well-being. However, existing research has revealed numerous barriers to personal sensing acceptance. This paper explores the influence of explanations on the acceptability of recommendations based on personal sensing. We conducted a qualitative study using five plausible personal sensing scenarios to elicit prospective users' attitudes towards personal sensing, followed by a reflective interview. Our analysis formed six nuanced design considerations for personal sensing recommendation acceptance: user personalisation, appropriate phrasing, adaptive capability, users' confidence, peer endorsement, and sense of agency. Simultaneously, we found that the availability of an explanation at each personal sensing layer positively influenced the willingness of the participants to accept personal sensing technology. Together, this paper contributes a better understanding of how we can design personal sensing technology to be more acceptable.