By Hamza Faisal

Spotify is done pretending it doesn't have a slop problem. The streaming giant announced on Tuesday that it has pulled more than 50 million AI-generated "slop and spam" tracks out of its recommendation engine so far this year — the largest single cleanup of synthetic music on any streaming platform to date.

The purge is not a one-time sweep. Spotify says it overhauled its recommendation system over the past six months, and the 50 million figure counts tracks demoted out of algorithmic playlists and personal mixes rather than deleted from the service entirely. On top of that, the company says it has removed more than 75 million tracks outright for serious or repeated violations of its rules — the same number it reported when it first announced its AI-music policies in September 2025.

The skip button became the first line of detection

Rather than betting everything on AI-detection software, Spotify turned its listeners into the first line of defense. Earlier this year, the company took a close look at the tracks people were skipping, ignoring, or telling the app they didn't want to hear. The pattern was unmistakable: the unwanted tracks were consistently the ones flagged as slop or spam.

That listener behavior now feeds a detection stack that layers engagement signals over upload patterns, track metadata, and several third-party AI detection tools. The company says the combination catches junk far more reliably than any single signal, and it hunts for known spam tactics: mass uploads, duplicates, SEO-stuffed titles, spammy playlist networks, and artificially short tracks built to game per-stream payouts.

The company reports a marked drop in listener complaints about slop and spam showing up in recommendations since the changes began about six months ago — though it offered no hard numbers for the improvement.

The money problem behind the junk

Spam tracks aren't just an annoyance; they directly drain musicians' wallets. Spotify pays royalties from a shared pool divided by each artist's proportion of total streams. When uploaders flood the catalog with synthetic songs — generated by the hundreds or thousands and looped with bot plays — total stream counts inflate and the slice going to human creators shrinks.

Beatdapp chief executive Morgan Hayduk put it bluntly: flooding catalog systems with synthetic tracks inflates total stream counts, directly shrinking the payout pool available for human artists. Spotify's 2024 rule that a track needs 1,000 streams in a year to earn anything was one early attempt to starve the scheme; demoting the tracks from recommendations is a much bigger lever.

The scale of the problem is industry-wide. Deezer reported that on peak days in June, more than half of everything uploaded to its platform was fully AI-generated, and up to 85 percent of those tracks' 2025 plays were flagged as fraudulent.

The problem Spotify can't quite fix yet

For all the momentum, Spotify is careful to hedge. The company admits that AI detection tools have real limits — and that as generative AI improves, identifying slop will only get harder. Critics note the 75-million-track deletion figure is unchanged from September 2025, raising questions about how much enforcement has actually happened since.

Still, the direction of travel is clear. Spotify now tags AI-created acts with an AI Persona badge that keeps them out of algorithmic recommendations unless a listener explicitly follows them, and it's beta-testing an Artist Profile Protection feature that lets verified musicians approve any new track uploaded under their name before it goes public.

The underlying message to the music industry is hard to miss: the age of letting machine-made filler silently siphon off streaming royalties is ending — at least on Spotify.

By Hamza Faisal