Spotify says it has excluded more than 50 million tracks it calls “slop and spam” from recommendations this year. The company announced the figure in an October 20 update on its artist protections. Those tracks were taken out of recommendations; Spotify did not say that all 50 million were deleted from its music catalog or generated by AI.

The distinction matters to listeners and artists. A track can remain available to search for or play while receiving less promotion through Spotify's recommendation system. The update gives a large company-reported count for that decision, but no public breakdown of the tracks, the assessment threshold or the effect on legitimate artists whose music might be caught by the filter.

The big change

  • What changed: Spotify proposed a music spam filter last year. It now says updates made over the past six months kept more than 50 million tracks it identifies as slop or spam out of recommendations in 2026. This is a claim about discovery, separate from removing tracks from the service.
  • Why it matters: Recommendation placement directs attention. Spotify is using listening behavior, upload patterns, metadata and outside AI detection tools together to decide which tracks should appear there. For artists, an erroneous classification could affect discovery even if a song remains playable; Spotify has not published a false-positive rate for this system.
  • What to watch: Spotify says it will keep changing the signals as upload tactics change. Public information about how artists can understand or challenge a recommendation exclusion would make the reported count more useful for judging the policy's effects.

Two different removal claims

The new Spotify statement describes more than 50 million slop and spam tracks taken out of recommendations “this year alone.” It separately refers to more than 75 million spammy tracks removed from Spotify altogether for serious or repeated rule violations. The first number concerns what the service recommends; the second concerns catalog removal. They should not be added together as a count of unique tracks, because Spotify does not explain whether the groups overlap.

The 75 million figure is also not a newly established one-year increase. In its September 2025 policy announcement, Spotify already said it had removed more than 75 million spammy tracks in the preceding 12 months. The October 2026 post repeats the threshold without a new period or a revised count. Neither announcement says that every track in either group was AI-generated.

In 2025, Spotify said its planned filter would tag tracks and uploaders engaged in tactics such as mass uploads, duplicate submissions, search manipulation and abuse of artificially short tracks, then stop recommending them. It said it would introduce the filter cautiously to avoid penalizing the wrong uploaders. Its new update reports multiple changes over the past six months, but does not specify when each took effect.

How Spotify says it chooses tracks

Spotify says it reviewed tracks that listeners skipped, ignored or explicitly rejected and found that tracks it identified as slop or spam appeared repeatedly in that group. It now combines listener engagement and feedback with upload patterns, metadata and information from several third-party AI detection tools. Spotify also says its editors do not program material identified as slop or spam. These are the company's descriptions of its process, not an independent measurement of how well it works.

Spotify acknowledges that AI detection has limitations and expects the task to become harder as generative tools change. It does not disclose the detectors, their accuracy, the weighting of the other signals, its definition of “slop,” or an independent audit of the 50 million figure. Spotify says listener concerns about this material in recommendations have fallen since the changes; the announcement supplies no counts or method for that comparison.

Other controls in the same update answer different questions. Spotify says its AI Credits let artists disclose which roles or instruments involved AI. Its AI Persona badge concerns an artist's presented identity, and those profiles are excluded from recommendations by default unless a listener follows one, for example. Spotify's earlier explanation of that badge says it describes the identity behind a profile, not how the music was made. These disclosure and identity measures do not turn the recommendation figure into a count of AI-made songs.

The platform choices also differ. Deezer has proposed a listener-controlled search toggle for fully AI-generated music, while YouTube Music is rolling out a player label for tracks with significant generative AI involvement. Spotify's reported 50 million figure concerns its own recommendation decisions about what it calls slop or spam. It does not describe a listener setting or a label on every affected recording.

For listeners, Spotify's account suggests fewer tracks it flags as spam in algorithmic discovery, while leaving the actual listening result unmeasured in the public post. For artists, the unresolved question is how accurately the service distinguishes manipulative uploads from legitimate music. The reported scale makes that question consequential even though the company has not released the evidence needed to answer it.

Sources & further reading