Beyond System Features: The Influence of Perfect Automation Schema on Trust Behaviors

Abstract

Human-machine interactions are becoming increasingly multifaceted. Appropriate human trust in these contexts is critical, given these systems’ capabilities and opacities. Behavioral proxies of trust provide a lens to infer psychological trust while quantifying human behavior directly. However, it remains important to investigate the role of users’ individual differences in human-machine interaction on behaviors even after the impact of system features are accounted for. One such individual difference is perfect automation schema, a construct which comprises high expectations and all-or-none thinking. In two experiments, we explored the influence of these factors on trust-relevant behaviors in human-machine interactions. Results show factors of perfect automation schema accounted for unique variance in users’ behaviors beyond manipulations. Some patterns of results (mis)align with postulates from the Human Factors literature, and the nuance of context and machine referent are discussed for future research.