German musical artist Holly Herndon just released an artificial intelligence tool of her own creation: Holly+. Herndon refers to Holly+ as her “digital twin” and for good reason. The digital voice instrument allows users to upload single- or multi-voice music and returns that music now sang by Herndon’s voice. Holly+ uses deepfake technology, trained by Herndon’s voice, to replicate her singing onto any audio file.
Although the technology is not new, Holly+ is technically the first tool of its kind. At present, there is no other readily available tool that allows for the recreation of audio files. Of course, the tool can only replicate songs to Herndon’s voice, but it may lead other artists to create their own “digital twin.”
Herndon made the tool in collaboration with Never Heard Before Sounds over the past year. She wrote on her blog that her motivation for creating this tool is rooted in “the popularity of celebrity vocal deep fakes already found all over YouTube.” She continued to write that she believes “that there will be demand for official, high fidelity, vocal models of public figures” as digital voice models become more advanced.
Herndon allowing and encouraging people to recreate music in her voice raises the question of who can claim the rights to the music. There is legal precedent that protects the appropriation of their vocal likeness, but there is no precedent for commercial exploitation via deepfake technology. It is currently assumed that Herndon will retain complete rights over any and all commercial exploitation created by Holly+. Herndon ultimately plans “to decentralize access, decision making, and profits made” from Holly+ and will eventually auction approved works of Holly+.
No comments:
Post a Comment