Spotify says it has taken more than 50 million slop and spam tracks out of recommendations this year. Its October 20 update puts a number on something you might have felt while listening: finding music now includes filtering out a lot of material made to occupy the feed.
The company separately reports removing more than 75 million spammy tracks from Spotify altogether for serious or repeated rule violations. Those figures describe different actions. Add them together and you'd invent a total the announcement doesn't establish.
What Spotify is counting
Spotify's recommendation figure covers slop and spam, rather than every song made with AI. It combines listener feedback with upload patterns, metadata and outside detection tools. "No single signal catches slop or spam on its own," the company writes, acknowledging that AI detectors have limits.
That's a sensible admission. A policy that depended on identifying the sound of AI perfectly would be a fragile way to run a music service. Looking at how tracks arrive and how people respond gives Spotify other evidence to use. The trade-off is that listeners can't infer the reason for a decision simply by hearing a track.
The numbers are Spotify's account of its own enforcement. They don't tell you how many mistakes it made, how decisions were appealed, or how much unwanted material is still reaching listeners. Those would be useful measures in the next update. A big removal count tells us the workload is enormous; it doesn't grade the decisions.
The labels answer different questions
AI Credits, visible in mobile Song Credits, describe disclosed AI contributions to a recording. Spotify's separate AI Persona badge concerns the identity presented by an artist profile. That August policy excludes AI Personas from recommendations by default, with exceptions such as following one yourself.
You can be interested in either question. Perhaps you enjoy a human musician experimenting with generated instruments but dislike being sold an invented person's biography. A single yes-or-no AI label would hide that difference. Credits and identity labels give you a more useful starting point, provided they're easy to find.
Spotify also says more than 300,000 artists are now verified. Verification identifies a real person behind the music; it shouldn't be read as a promise that no AI tools touched the recording. Check the credits when that matters to you.
What listeners and musicians can do
Follow the artists you want to hear, look beyond the cover image when a new profile seems suspicious, and use the feedback controls when recommendations miss. For AI-assisted releases, decide what kind of involvement you're comfortable with. Our earlier ElevenLabs Music coverage explains why generating a track and having the rights to release it are separate questions.
If you're a musician, Artist Profile Protection lets eligible participants review deliveries that name them before those releases appear on their profile. Spotify says wider access is due by year-end. Check your artist account for availability, particularly if a release you didn't make has appeared under your name.
Spotify's strongest next move would be to show listeners why a recommendation earned its place. Removing millions of tracks is a start; making the remaining ones worth your time is what you'll hear.
Cover: AI-generated editorial illustration by TMRW using Spotify branding and its October 20 reported recommendation count.
