How do Blame Attributions Impact Trust in Complex Task Environments?

Abstract

Trust in automation is a multidimensional process in predicting the proper use of automated systems. Trust can be generalized across components in multi-component systems, leading to contagion effects. However, no research has considered how blame attributions influence the size of contagion effects. We conducted an experiment to address how attributions of blame and responsibility influenced the magnitude of contagion effects for trust. Participants were presented with a hypothetical scenario depicting an airplane environment consisting of six human and six automation components. Participants rated their trust in each component before and after an oxygen mask failure. The findings indicate that the percentage of blame attributed to a component predicted the degree to that trust decreased following an error. Thus, more research should focus on how users assign blame across systems.