The need for specificity in uncertainty communication is further supported: Preliminary evidence from eye tracking

Abstract

Automation transparency is understudied in environments with varied automation reliability. We previously found uncertainty communication, a form of automation transparency, led to nuanced and unintentional performance outcomes in this type of environment. To better understand these outcomes, participants’ monitoring policies of automation were studied via a preliminary eye tracking analysis. In a simulated multi-UAV mission, 272 Naval Aviation trainees monitored three automated systems that were identical in function, but differed in reliability. Halfway through the experiment, uncertainty communication about the least reliable system was delivered as a chat message. Performance improvements coincided with a monitoring policy that had larger dwell times overall and one where the proportion of visual attention towards the automated system mentioned in the uncertainty communication increased upon its delivery. Overall, examining participant monitoring policies helped explain other performance outcomes and bolstered our previous conclusion: effective uncertainty communication requires maximum specificity.