Trust in decision support: eye-tracking as a continuous measurement

Decision aids are developed to ease cognitive load on operators interacting with complex automated systems; however, critical human components are often ignored during design. Finding an appropriate balance of automated assistance and operator trust is paramount in achieving optimal output from the human-automation interaction. Establishing a consistent metric of trust measurement will enhance the functional design of automated decision support, especially as the use of eye tracking opens the field to the use of real-time measurements. This study tasks participants to make measurements, assisted by a decision aid system, within a bone defect model image. The study tests for correlation between eye tracking data and participant trust survey answers. Results do not indicate a significant correlation; however, fixation duration and fixation count on decision aids rise as decision aid reliability decreases. These results support eye tracking's potential as a real-time, continuous measurement of human trust in automation.