Last week, Grimes gave creatives the go-ahead to use AI-generated versions of her voice in their art, offering a 50% royalty split and musing on the possibility of "open sourcing all art and killing copyright." On April 30, she shared a beta version of the GrimesAI-1 Voiceprint.
Users who upload their vocal files to the site will receive output audio generated in the style of Grimes' voice. “If you are signed to a label or wish to distribute it yrself," Grimes wrote on Twitter, "plz email us at firstname.lastname@example.org." She requested that the songs using her voice be "tasteful" — last week, Grimes said "rly rly toxic lyrics w grimes voice" such as "baby murder" songs or "Nazi anthems" will face takedown notices. 50% of the master royalty recordings from songs using the Grimes AI will be split between Grimes and the song creator.
Just prior to the debut of the AI, Grimes gave a keynote speech at the International Music Summit in Ibiza. With a conference theme of "Face The Future," Grimes expanded her thoughts on AI and dismissed copyright as repressive. “I feel strongly that there’s way too much gatekeeping in music,” said Grimes. “Copyright sucks," she said. "Art is a conversation with everyone that’s come before us. Intertwining it with the ego is a modern concept. The music industry has been defined by lawyers, and that strangles creativity.”
Last week, Grimes parted ways with her label Columbia Records. She signed with the company in 2021 and released two singles: “Player of Games” and “Shinigami Eyes.” The AI site contains these two songs along with “Passage (Out) - demo,” a new song incorporating Plastikman's 1998 song with a spoken word manifesto from Grimes.