Spotify Removes 75M AI Tracks, Bans AI Voice Clones

Spotify says it has removed roughly 75 million “spammy” AI-generated tracks and unveiled new policies to combat AI-driven spam, impersonation and deceptive uploads that it says siphon royalties and degrade listener experience. Measures include a ban on unauthorized AI voice clones, a music-spam filter to flag mass uploads, duplicates, SEO-hacked metadata and artificially short tracks, and faster content-mismatch review that lets artists report problems even pre-release. Spotify is adopting DDEX metadata standards for voluntary AI-use disclosures so creators can credit AI in vocals, instrumentation or production, and says more than 15 labels and distributors have committed. The announcement follows high-profile abuses on the service, including AI-generated Velvet Sundown and fake Volcano Choir uploads, which executives say helped prompt the changes. Spotify says it will roll out filters cautiously and continue refining detection signals to avoid unfairly penalizing legitimate creators while permitting responsible AI-created music.