Abstract
Cognitive bias can prompt inconsistency and error in visual comparisons of forensic patterns. We tested whether bias can likewise impede attempts to identify unknown criminal perpetrators via facial recognition technology (FRT). Participants (N = 149) completed two simulated FRT tasks. In each, they compared a probe image of a perpetrator's face against three candidate faces that FRT allegedly identified as possible matches. To test for contextual and automation biases, each candidate was randomly paired with either extraneous biographical information or a biometric confidence score, respectively. As predicted, participants rated whichever candidate's face was paired with guilt-suggestive information or a high confidence score as looking most like the perpetrator's face, even though those details were assigned at random. Furthermore, candidates randomly paired with guilt-suggestive information were most often misidentified as the perpetrator. These findings indicate a clear need for procedural safeguards against cognitive bias when using FRT in criminal investigations.