Leveraging gaze for potential error prediction in AI-support systems: An exploratory analysis of interaction with a simulated robot | Synapse