Abstract
Predictor displays are a form of automation that have been shown to provide a notable benefit in many environments and are applied to wildfire suppression in the present study. 22 participants were recruited to participate in a wildfire prediction study in which they predicted the fire trajectory based on the automated predictor and their interpretation of the raw data. In all scenarios, a large predictor benefit was observed, and in four of six scenarios, significant signs of automation bias were found, following an erroneous automated prediction. Participants performed significantly worse when the raw data influences on the trajectory were removed in a control study, signaling the value of such data in calibrating dependence on the predictor. The FIPRED display was beneficial to participants, improving their ability to predict the future position of the fire. However, the automation bias indicated that participants may have over-relied on the automated assistance.
Get full access to this article
View all access options for this article.