Autonomy-Respectful E-Coaching Systems: Fending Off Complacency

Abstract

Failing to act in line with one’s own standards–self-regulation failure–has been linked to problematic behaviors such as bedtime procrastination, educational underachievement, domestic violence, and substance abuse. Technological advancements in the collection and mining of personal data are leading to the development of adaptive, personalized systems that support individuals in their self-regulation through automated “e-coaching”. This trend raises the question: What are the risks of being supported in one’s self-regulation by such “e-coaching systems”? Answering this question is the focus of this dissertation. I lay the groundwork by elaborating my conception of self-regulation failure and proposing a novel definition of “e-coaching systems” that highlights the way these systems make suggestions for action that users will not have seen before. This insight brings into focus specific ethical concerns that I discuss in surveying the “ethical landscape” of e-coaching systems, which includes concerns about social justice, infringements of rights, and unintended negative consequences for people’s personal autonomy. In relation to this final category, I argue in the remainder of the dissertation that, given the way in which e-coaching systems operate–by means of persuasive techniques that may change over time and across contexts–there is a danger of users becoming complacent with regard to assessing their e-coach’s suggestions in relation to their own goals, plans, and values. This lack of effort, I argue, erodes one’s practical standpoint, thereby undermining the extent to which complacent users determine their own course in life. The dissertation concludes with a set of “anti-complacency” design recommendations.