The Voice and Eye Gaze Behavior of an Imposter: Automated Interviewing and Detection for Rapid Screening at the Border

Contextual differences present significant challenges when developing computational methods for detecting deception. We conducted a field experiment with border guards from the European Union in order to demonstrate that deception detection can be done robustly using context specific computational models. In the study, some of the participants were given a “fraudulent” document with incorrect data and asked to pass through a checkpoint. An automated system used an embodied conversational agent (ECA) to conduct interviews. Based on the participants’ vocalic and ocular behavior our specific model classified 100% of the imposters while limiting false positive errors. The overall accuracy was 94.47%.