Spotify said it has removed more than 50 million AI-generated “slop and spam” tracks from recommendations over the past year and adjusted its recommendation systems to favor human artists. Both Engadget and The Verge report that the company says listener complaints about such content have fallen markedly since the changes.
Spotify says it identifies problem tracks using metadata, user feedback, upload patterns, and third-party AI detection tools. It has also introduced an “AI Persona” label for virtual artists, and those tracks are excluded from editor-curated playlists and default recommendations.
However, the two outlets differ in emphasis. Engadget points out that Spotify’s overall removals for policy violations still stand at 75 million tracks, the same figure the company cited in September 2025, and that Spotify has not offered users a way to filter out AI audio. The Verge does not raise those concerns, instead noting that rivals such as Deezer, Bandcamp, and YouTube have adopted stricter policies on AI music.
Spotify is also leaning into generative AI, partnering with Universal Music Group and Believe on AI remix tools. That leaves the company trying to curb spam while building AI features into its platform, a balance that critics say remains unresolved.