Spotify purged 50M AI slop tracks to save human music

If you have used Spotify lately and felt like your recommendations were slowly getting flooded by soulless, automated background noise, you are not imagining things. The platform has spent the last year fighting off a massive wave of synthetic filler.

Spotify shared an update on its efforts to protect human artists and clean up its catalog. The numbers are staggering: so far this year, the company has booted more than 50 million AI-generated slop and spam tracks out of its recommendation feeds.

On top of that, Spotify completely banned and deleted over 75 million spammy tracks from the platform for severe or repeated rule violations.

Flushing out the AI audio junk

Earlier this year, Spotify analyzed the tracks users were constantly skipping, ignoring, or reporting. The pattern was unmistakable: the vast majority were AI-generated spam or mass-produced slop.

Over the past six months, engineers made widespread updates to the recommendation system to suppress low-quality junk and prioritize human creators. Editorial playlists follow the same guidelines, and user complaints about bad recommendations have dropped noticeably.

Catching this content is getting harder, though. Spotify admitted that AI detection tools alone are not enough and will only become less reliable as generative tools improve.

Instead, the platform relies on a combination of user engagement, upload frequency patterns, metadata analysis, and multiple third-party AI detectors. They are actively hunting down mass uploads, duplicates, SEO keyword gaming, coordinated playlist networks, and artificially short tracks designed to game streaming metrics.

Five guardrails for human artists

Beyond backstage cleanup, Spotify has rolled out several public-facing tools to protect real creators from being crowded out.

First is AI Credits, launched in April, which now logs tens of thousands of submissions every day from artists disclosing exactly where generative tools were used. Listeners can inspect these in the mobile song credits view.

Second is the Verified by Spotify badge, now covering over 300,000 artists who account for more than 99% of total searches on the service, with more than 70% of them being independent creators.

Third, fully synthetic acts are now tagged with an AI Persona label. These accounts are excluded from editorial and algorithmic recommendations by default unless a listener explicitly follows them. Spotify claims to be the only major streaming platform doing this.

Fourth, a new Artist Profile Protection feature gives musicians final approval over any new release delivered under their name before it hits their profile. A few thousand artists are testing the beta now, with a platform-wide rollout slated before the end of the year.

Finally, the 1,000-stream royalty threshold introduced in 2024 has become a structural barrier. Tracks showing strong AI signals are 10 times more likely to fall below this floor, making high-volume spam campaigns economically unviable.

Why it matters

For regular listeners, this cleanup keeps your Discover Weekly from turning into an AI-generated elevator music swamp. For independent musicians, the upcoming profile protection tool finally offers a real defense against spoofing and identity theft.

We think Spotify's strategy here is clever because it attacks the problem through economics rather than moral posturing. Instead of attempting an impossible outright ban on generative audio, Spotify cut off the two things bad actors care about: algorithmic reach and royalty payouts.

By ensuring tracks flagged as AI slop are 10 times more likely to miss the 1,000-stream payout threshold, churning out millions of low-effort tracks suddenly stops being profitable. Locking AI Personas out of recommendations by default achieves the same thing for user attention.

Still, the cat-and-mouse game will only get tougher. As generative models sound less like synthetic mush and more like polished studio productions, third-party detection signals will become shakier. Spotify will have to be careful not to penalize legitimate human creators experimenting with new production tools.

This also puts serious pressure on rivals like Apple Music and YouTube Music. If other platforms do not adopt similar disclosure tags and algorithmic quarantines, they risk becoming dumping grounds for the millions of spam tracks Spotify is currently rejecting.