Human interaction with automated aids: Implications for robo-advisors

Abstract

Robo-advisors, a type of automated decision aid, offer consumers a cost-efficient alternative to traditional financial advisory services. Because aids do not always produce correct judgments, however, users may fail to act appropriately on their advice. To anticipate and protect against suboptimal aid use, designers need to understand the variables that influence automation trust and dependence, including the operators' inherent biases, and the characteristics of the automated system itself. This paper reviews the literature on human interaction with decision aids, aiming to inform the design of robo-advisory platforms.