Artists Rage as Suno AI Training Database Exposed
Get the Culture newsletter
Daily culture — film, music, books, the trends and ideas worth your attention. Free.
- Suno admitted in a legal filing responding to an RIAA lawsuit from Universal, Sony, and Warner that it trained its model on "tens of millions of recordings" available online, a disclosure that producer Mike Sabath called "crazy."
- The Atlantic released the Laion-Disco-12M dataset — 12 million songs pulled from YouTube spanning more than 250,000 artists — and SZA found 238 of her songs listed while Kehlani found 179, prompting both to blast the practice on Instagram.
- SZA argued that "the best and brightest Black minds of writers and producers" are being used to train AI on platforms like Suno, a racial dimension to the training debate the broader coverage has largely overlooked.
- Doja Cat revealed that a heap of tracks fans believed to be unreleased leaks were actually AI-generated, with some followers still refusing to believe her, eroding the trust contract between artists and audiences.
- TIDAL announced on June 29 that it will tag music found to be 98–100% AI-generated starting in mid-July and block it from monetization, while Spotify and Apple Music rely on artist self-disclosure for similar labels.
- Second Voice's AI-generated hit "Let Me Be" — built by Rwandan artist Elvin Cena using Suno — spent six weeks at No. 1 on Billboard's U.S. Afrobeats chart alongside Tyla, Ayra Starr, and Asake, illustrating Jen Jacobsen's warning that stolen work is now "competing directly" with the artists it was taken from.
- The Artist Rights Alliance, whose April 2024 open letter drew 200+ signatories including Billie Eilish and Chappell Roan, says the issue has shifted from copyright infringement to "participation in the marketplace" — artists need transparency before they can demand anything else.
Why it matters: The Atlantic's searchable database turned a theoretical grievance into a personal one for more than 250,000 artists, and TIDAL's mid-July tagging policy is the first concrete industry enforcement mechanism — one that blocks AI-only releases from earning royalties while leaving 50–97% AI-assisted tracks untouched. Artists lose both income and identity when fans can no longer tell real from fake.
