What An Anonymous Artist Taught Us About The Future Of AI In Music

Hidden behind a white sheet and sunglasses, an anonymous music-maker called Ghostwriter logged on to TikTok in April to post his first video. In it, he announced his debut single, “Heart on My Sleeve” ­— and signaled a seismic shift in the music business. Using an artificial intelligence voice filter that disguised Ghostwriter’s own timbre behind that of Drake and The Weeknd, “Heart on My Sleeve” was the first song to show just how far AI music had come already and what novel challenges and opportunities it would present to artists.

The song also proved that artists would struggle to control how their voice and likeness are used in the age of AI ­— even superstars like Drake and The Weeknd who have the resources to fight back. Universal Music Group (UMG) issued a strongly worded statement soon after the song’s posting, condemning “infringing content created with generative AI,” and the song was quickly taken down from most platforms. But fans continued to post it to TikTok and YouTube every time it was removed. As Ghostwriter later put it in his Billboard cover story: “The genie can’t be put back in the bottle.”

“Heart on My Sleeve” forced the industry to reckon with the limitations of existing “right of publicity” laws that protect artists from having their voices and likenesses commercially exploited without their authorization. The strength of this right varies from state to state, and at the time the song was released, there wasn’t a precedent for issuing takedown notices for these types of violations, like there is for copyright infringement claims.

Likely as a result of the song and other AI concerns, UMG general counsel/executive vp of business and legal affairs Jeffrey Harleston went to Capitol Hill to talk to the Senate Judiciary Committee in July, asking for a “federal right of publicity” to be created to help protect artists. Streaming services also voluntarily entered talks with the major labels to form a system so that labels can request takedowns for right of publicity concerns.

Still, some artists consider Ghostwriter to be a revolutionary. Artist-producer Grimes took to X (formerly Twitter) after the release of “Heart on My Sleeve” to write, “I think it’s cool to be fused w a machine, and I like the idea of open sourcing all art.” Shortly after, she launched GrimesAI, a voice model trained on her recordings that lets fans shape-shift their voice into hers at the click of a button. She also collaborated with TuneCore to allow fans to distribute their resulting songs to streaming services under the tag “GrimesAI” to distinguish it from her own catalog.

YouTube recently built on this same idea with the experimental program Dream Track. It lets fans use voice models of Charlie Puth, John Legend, Sia, T-Pain, Demi Lovato, Troye Sivan, Charli XCX, Alec Benjamin and Papoose with the approval of partners UMG and Warner Music Group (WMG).

Songwriters and publishers have also started using AI voice technology. This summer, Billboard reported that AI voice models were used to help writers pitch song ideas to artists. The AI-altered pitch records had been used as an internal reference for the writers to gauge if the track sounded appropriate for a particular singer while composing, and sometimes the records with the AI voice applied were also sent to the artist’s team in hopes that they would increase their chance of landing the song.

In a recent interview with Billboard, pop artist Lauv said he has seen that usage firsthand with one of his songwriter friends who was pitching a track to Nicki Minaj using AI. Lauv himself is also leaning into the new technology: In November, he partnered with AI voice startup Hooky — one of many startups in the space — to translate his new single, “Love U Like That,” into Korean. After getting Korean American songwriter Kevin Woo to translate the lyrics and sing a Korean version, Hooky applied Lauv’s AI voice model to map Lauv’s signature sound over Woo’s.

Ghostwriter’s manager previously said he also believes that artist estates and catalog owners could use AI voice models to market their music. WMG is now experimenting with the catalog of late French singer Édith Piaf. The company is creating AI models to resurrect her voice and likeness for a controversial new biopic called EDITH to “further enhance the authenticity and emotional impact of the story,” according to a press release.

“Heart on My Sleeve” represents just one of many burgeoning areas of AI that can transform the music business in the future. Concurrently, as startups like Kits.AI, mayk.it, Hooky, Voice-Swap, Supertone and Lingyin Engine hope to capture the AI voice model market, AI-generated beats and loops promise to upend production music libraries and the overall creation process. AI “functional music,” soundscapes designed to fit user needs for sleep, focus or relaxation, also promise an evolution in wellness and ambient music. And as AI stem separation unlocks possibilities for remixing, synch licensing and restoring old audio files (the new Beatles song, “Now and Then,” used the technology to restore John Lennon vocal recordings), there are surely even more use cases for AI that are yet to be discovered.

While legal and logistical questions remain with the technology, Ghostwriter’s team is pressing on. He has teased a second single (“Whiplash” featuring the AI voices of Travis Scott and 21 Savage) while expressing his desire to collaborate with other artists. As such, Ghostwriter’s manager believes it is of the utmost importance for the music business to band together to “define an equitable arrangement for all stakeholders” when using AI. As he previously told Billboard, “We had an opportunity [with “Heart on My Sleeve”] to show people the value in AI and music … I like to say that everything starts somewhere.” 

This story originally appeared in the Dec. 9, 2023, issue of Billboard.