Trust in Automation: The Effects of System Certainty on Decision-Making

Abstract

In the field of Human Factors, the concept of trust in automation can help to explain how and why users interact with particular systems. One way to examine trust is through task performance and/or behavioral observations. Previous work has identified several system-related moderators of trust in automation, such as reliability and complexity. However, the effects of system certainty, i.e., the knowledge that a machine has regarding its own decision-making abilities, on trust remains unclear. The goal of this study was to examine the extent to which system certainty affects perceived trust. Participants performed a partially simulated flight task and decided what action to take in response to targets in the environment detected by the aircraft’s automation. The automation’s certainty levels in recognizing targets were 30%, 50%, and 80%. Overall, participants accepted the system’s recommendation regardless of the certainty level and trust in the system increased as the system’s certainty level increased. Results may help to inform the development of future autonomous systems.