YouTube's "approach to responsible AI innovation" is greenlighting the widespread posting of "synthetic media." While major-label musicians' requests for video-takedown might be being taken seriously by the platform, all that its statement says about deepfakes is that the company will "consider a variety of factors when evaluating" such reports. Youtube needs to make much more convincing reassurances.
Google Deepmind — the UK-based AI research lab that has launched YouTube's Dream Track — has already reassured those concerned about deepfakes and copyright that any synthetic audio it publishes will be watermarked. This watermarking will be "inaudible to the human ear and doesn't compromise the listening experience," but allows for detection even after manipulation. This fact ought to lay to rest any fears among skeptics.