Perhaps the most insidious security challenge is not technical but psychological. Judges and juries are conditioned to trust what they can see and hear, but deepfakes exploit this intuition. A fabricated video showing a defendant committing a crime can be extraordinarily persuasive. At the same time, the existence of deepfakes makes it easy for guilty parties to dismiss genuine recordings as fabrications, a dynamic Chesney and Citron (2019) capture in the concept of the “liar’s dividend.” In this sense, deepfakes threaten to destabilize one of the pillars of evidentiary security: confidence in audiovisual materials as reliable truth.