When people experience the same automation, their trust in automation can diverge. Prior research has used individual differences—trust propensity and complacency—to explain this divergence. We argue that bifurcation as an outcome of a dynamic system better explains trust divergence. Linear mixed-effect models were used to identify features to predict trust (i.e., individual differences, automation reliability, and exposure). Individual differences associated with trust propensity and complacency increases the R2 of the baseline model by 0.01, from R2 = 0.40 to 0.41. Furthermore, the Best Linear Unbiased Predictors (BLUPS) for random effect of participants were uncorrelated with trust propensity and complacency. In contrast, modeling trust divergence from a dynamic perspective, which considers the interaction between reliability and exposure along with the individual by-reliability variability fit the data well ( R2 = 0.84). These results suggest dynamic interaction with automation produce trust divergence and design should focus on state dependence and responsivity.