4 ms·
Looking at the video here, the two best by far are the Michael Caine and Julie Andrews deep fakes, and in both cases the voice is doing most of the heavy liftin
by fullshark 6y ago
Looking at the video here, the two best by far are the Michael Caine and Julie Andrews deep fakes, and in both cases the voice is doing most of the heavy lifting. Deepfake audio is somewhat scarier to me in terms of political / legal chaos than video, much easier to possibly trick someone with a "secret" audio recording than video recording if we some day get to the point of near identical audio mimicry.
- DerDangDerDang 6y agoNot sure there’s any deepfakery on the Michael Caine audio. Peter Serafinowicz (the other collaborator and originator of ‘sassy Trump’) is well known for his Michael Caine impersonation.
- filoleg 6y ago>Not sure there’s any deepfakery on the Michael Caine audio Most likely not, and the parent comment seems to agree with you on that. I think they were just trying to point out, in general, that the arrival of commonplace audio deepfakes might be way more disruptive than video deepfakes, despite a lot of people (including myself) who used to counter-intuitively think that video deepfakes would be more disruptive.
- DerDangDerDang 6y agoFair point, and I agree audio is potentially more disruptive - especially if it can get to good real-time performance.
- lillesvin 6y agoAudio deepfakes are getting better but they have a surprisingly long way to go still. Here's an exploration of a deepfaked Jay-Z reading/rapping the Navy Seal copypasta: https://www.youtube.com/watch?v=UZzYoOdIXoQ https://www.youtube.com/watch?v=UZzYoOdIXoQ