Transparency and fairness. French music streaming platform Deezer has developed technology that allows it to identify songs that clone the voices of music stars via artificial intelligence, it announced Tuesday.
"Our goal is to eliminate illegal and fraudulent content, increase transparency and develop a new remuneration system where professional artists are rewarded for creating valuable content," Deezer CEO Jeronimo Folgueira said in a statement.
"That's why ... we are developing tools to detect AI-generated content," he continues. This so-called marking system aims first to identify songs using "synthetic voices of existing artists".
" READ ALSO "Inflate your listening is to steal honest artists": how Deezer tracks fake streams
"This information will be used to notify artists, labels and users of AI-generated content on the platform," Folgueira said. For him, this new system intends to "develop a remuneration model that distinguishes between the different types of musical creation".
An AI like Eminem's voice
An announcement that comes at a time when artificial intelligence is booming and shaking up many sectors, including music.
In mid-February, the star DJ David Guetta had made talk about him by using AI for a voice like rapper Eminem for one of his shows. The artist did not market this title, explaining to the BBC that he wanted to "open the discussion for awareness".
No, Eminem did not feature David Guetta. On February 3, he shared images of him playing a song with the voice of the interpreter of "Lose Yourself". pic.twitter.com/ivkPYe6Ugd
— HuffPost (@LeHuffPost) February 9, 2023
At the heart of the debate is a question: does intellectual property exercise on one voice? "What you protect with copyright is the expression of an idea, and the voice is not really that," Andres Guadamuz, who teaches intellectual property law at Britain's University of Sussex, said in May.