The Effects of Individual Differences in Working Memory on Trust Recovery

Abstract

While trust repair is an active area of inquiry (i.e., systems that proactively repair users’ trust in itself, De Visser, Pak, and Shaw 2018), most technology does not have the ability to initiate trust repair. Thus, it is crucial to examine how unaided trust recovery occurs and the factors that could influence it. Prior research has highlighted the potential importance of individual difference factors such as working memory capacity (e.g., Rovira, Pak, McLaughlin, 2017). Previous research also suggests operators’ trust on automation recovers naturally without trust repair strategies (e.g., Lee and Moray, 1992; Mishler and Chen, 2023). However, the specific dynamics of unaided trust recovery and how they are influenced by individual differences are unknown. There are theoretical and practical reasons why this topic is important. Theoretically, it is important to know why some people show trust recovery (i.e., trust recovery naturally, over time), but not others, as such information can inform theories of human-machine trust. Practically, a lot of technologies do not have a trust repairing system to rebuild users’ trust after errors occur; information gained in this study may have direct implications for redesigning systems to aid trust recovery (without using trust repair). The current study has two goals; the first is to understand how trust recovers naturally (unaided). The second goal is to specifically examine how working memory is related to the process of unaided trust recovery. This study will use a computer-based automation task to examine participants’ trust level at three different time points (i.e., before, during, and after automation failures). A final exploratory role of this study will be to examine the role of attention control on trust dynamics. The current study will shed light on individual differences in unaided trust recovery and potential causes.