Abstract

Deepfakes are extremely realistic audio/video media. They are produced via a complex machine-learning process, one that centrally involves training an algorithm on hundreds or thousands of audio/video recordings of an object or person, S, with the aim of either creating entirely new audio/video media of S or else altering existing audio/video media of S. Deepfakes are widely predicted to have deleterious consequences (principally, moral and epistemic ones) for both individuals and various of our social practices and institutions. In this introduction to the Topical Collection, I first survey existing philosophical research on deepfakes (Sects. 2 and 3). I then give an overview of the papers that comprise the Collection (Sect. 4). Finally, I conclude with remarks on a line of argument made in a number of papers in the Topical Collection: that deepfakes may cause their own demise (Sect. 5).