ARTICLE: Grimes says anyone can use her voice for AI-generated songs.

Grimes has other plans than those of the major record labels, who have tried to stop him from using AI voice tracks of well-known artists in his songs.

Grimes wrote on Twitter late Sunday night, “I’ll split 50% profits on any successful AI-generated song that utilizes my voice. She said, claiming she has no label and “no legal restrictions,” “Feel free to utilize my voice without penalty.”

At this moment, Grimes’ Sunday night tweet appears to be nothing more than a late-night tweet with the potential to develop into something later. Grimes said the profit sharing may apply to “viral” or “very popular” music that currently exist but didn’t go into much detail about how such agreements would operate.

Grimes is hardly the first musician to use artificial intelligence and voice cloning techniques. In 2021, experimental musician Holly Herndon debuted Holly Plus, her own synthetic voice. Users of Herndon can contribute audio files to get a new one sung in her voice. The voice model is only available to participants in Herndon’s decentralized autonomous organization (DAO).

More readily available than ever, voice models trained on a corpus of recordings of an artist produce strange, amusing, and occasionally unsettling outcomes. Last week, a song created using voice models for Drake and The Weeknd suspiciously went viral before being swiftly and uniformly removed from streaming services. In the midst of the song “Heart on My Sleeve” going viral, Universal Music Group released a scathing statement in which they claimed that copyright violations were being committed by training AI models on the works of its artists.