ABSTRACT
This chapter offers a definition of deepfakes and explains the technologies that make them possible. We then discuss a set of ethical and epistemological issues, based on research from philosophy, technology and computer studies, and scholarship on the ethical and legal regulation of deepfake technologies. First, we examine an existing worry about the impact deepfakes could have on our ability to trust videos. As deepfakes become more prevalent in our online environments, the worry is that we will place less trust in the videos we watch online, relying less on videos to verify real-world events. Second, we consider the extent to which deepfakes might undermine the amount of information videos transmit to their audiences. The worry here is that deepfakes will become increasingly sophisticated, and so increasingly difficult for to distinguish from genuine videos. In turn, the information carried by a genuine video is at risk of being undercut by a sophisticated deepfake. The upshot is that we either suspend judgement or take the risk of acquiring false beliefs. Finally, we turn our attention to a number of understudied possibilities, concerning the relation or deepfakes to epistemic luck, epistemic rights, and epistemic character. The proliferation of deepfakes arguably increases the risk of veritic epistemic luck and jeopardises several of our epistemic rights, such as our right not to be deceived. We close by relating deepfakes to issues of epistemic character, focusing on epistemic kinds of cynicism and insouciance and what Shannon Vallor calls ‘technomoral virtues’.
