Spotify says it removed millions of 'spammy' AI-generated music tracks

Spotify has removed more than 75 million AI-generated “spammy” music tracks from its platform over the past 12 months, the company said Thursday, stepping up its crackdown on unauthorized AI-generated use of artists’ voices.

The Swedish audio company said it would improve its enforcement of impersonation violations, launch a new spam filtering system and work with partners to label tracks that incorporate AI.

“We envision a future where artists and producers are in control of how or if they incorporate AI into their creative processes,” Spotify said in a post on its website on Thursday. “As always, we leave those creative decisions to artists themselves while continuing our work to protect them against spam, impersonation, and deception, and providing listeners with greater transparency about the music they hear.”

The push comes as tech platforms are grappling with how to handle the significant increase in AI-generated content. While some creators have embraced the new tech tools, others say their businesses are hurting from people who have used AI to impersonate them without their permission.

“As far as what Spotify is doing, I think they’re doing the right thing for the artists and doing the right thing to preserve Spotify’s integrity at this point,” said Rob Enderle, principal analyst of Bend, Ore.-based advisory services firm Enderle Group.

Spotify said it would only allow vocal impersonation in music if the artist authorized it and plans to reduce the wait time for review on content mismatches, “enabling artists to report ‘mismatch’ even in the pre-release state.”

“Unauthorized use of AI to clone an artist’s voice exploits their identity, undermines their artistry, and threatens the fundamental integrity of their work,” Spotify said. “Some artists may choose to license their voices to AI projects — and that’s their choice to make. Our job is to do what we can to ensure that the choice stays in their hands.”

The growing popularity of AI tools will increase the proliferation of online content generated by the technology including deepfakes, making it more difficult for tech companies to police, experts say.

“We’re at the very beginning of this,” Enderle said. “The tools are becoming far more advanced as we speak, so this problem isn’t going to be getting smaller, it’s going to be getting quite a bit bigger, very quickly.”