Skip to main content

Timeline for Using deepfake to create endorsment

Current License: CC BY-SA 4.0

11 events
when toggle format what by license comment
Jun 17, 2022 at 19:28 vote accept Yanick Salzmann
Aug 18, 2021 at 19:15 comment added David Siegel @Yanick Salzmann To the best of my knowledge there are no special rules for voice imitations produced using an AI, the rules are the same as for using a voice imitator.
Aug 18, 2021 at 17:04 comment added Yanick Salzmann @gnasher729 Thats why I specifically asked about non-visual deepfakes only working on audio. I think the answer covers it and is actually quite simple: It is the same as when you are using a voice imitator. I had just naturally assumed that there might be different rules when its AI doing it, but I assume there are not.
Aug 18, 2021 at 16:31 history edited David Siegel CC BY-SA 4.0
furthe expand answer
Aug 18, 2021 at 16:27 comment added David Siegel @gnasher729 well he could in some US states. and some other places Not all jurisdictions recognize rights of personality or rights of publicity.
Aug 18, 2021 at 16:25 comment added gnasher729 @Dave, the technology is irrelevant. If you find someone who looks like the spitting image of Donald Trump, and show him on TV saying "Hi, I am Donald Trump, and I tell you that Joe Smith's potato chips are the best in the world", then Donald Trump can sue you.
Aug 18, 2021 at 16:24 comment added David Siegel @Dave I have added to the answer based on your comment, does it address the question better now?
Aug 18, 2021 at 16:23 history edited David Siegel CC BY-SA 4.0
expand answer in response to comment
Aug 18, 2021 at 16:14 comment added David Siegel @Dave. I see, I didn't quite understand the question. I am not at all sure what the status would be under copyright law. But my comments in the answer about rights of publicity and false advertising if this was done without permission still apply. The technology does not change those issues at all.
Aug 18, 2021 at 15:03 comment added Dave This does not seem to be answering the question asked. The BBC had an article on the technology. It is quite technically feasible to train an AI of publically available voice samples. If one did this without getting permission, would the sources of the training set have any claim on the output of the neural network, assuming it did not explicitly claim to be {performer}?
Aug 18, 2021 at 14:35 history answered David Siegel CC BY-SA 4.0