Spotify just pulled more than 50 million songs it called AI-generated “slop” or spam out of its recommendation system this year. It’s the biggest step the company has taken so far to deal with low-quality, mass-produced synthetic music.
They made the news official on October 20, saying that over the past six months, they’ve changed their systems to weed out these AI tracks from the playlists and suggestions you see—both the ones tailored for your tastes and the ones put together by their editors.
But don’t get the wrong idea: Spotify hasn’t deleted these 50 million tracks. You can still find them if you search for them specifically. They just won’t pop up in your Discover Weekly or other recommendations.
Spotify summed it up: “AI-generated slop and spam don’t belong in your recommendations, and we’ve made changes to keep them out and protect the listening experience.”
Separately, Spotify has actually taken down more than 75 million tracks for serious or repeated policy violations like spam and impersonation. That number is totally separate from the tracks they pulled from recommendations.
So, how do they spot this AI spam? Spotify uses a mix of things—how listeners behave, user feedback, how quickly and often tracks get uploaded, the metadata attached to the music, and third-party AI detection tools. Listeners skip or report these tracks a lot, which is a giveaway.
They’re careful with their language. Spotify isn’t saying every single excluded song is definitely AI-made. They’re talking about tracks that look AI-generated or match their idea of slop or spam.
To help clean things up, Spotify’s also launched an “AI Persona” label for accounts that use AI-created identities instead of linking to real artists. Music from those accounts gets left out of your recommendations by default.
Human artists now get verified, and there’s a new way for creators to share if they used AI in their song credits. But there’s still no rule saying every song that used AI at any stage needs to be labelled.
Why does any of this matter? Well, Spotify’s recommendation system drives music discovery. Playlists and algorithmic suggestions shape what people listen to. By filtering out AI “slop,” Spotify limits how far that stuff spreads—even though it’s technically still there if you go look for it.
For indie musicians, this is a win; it means less competition from auto-generated uploads flooding the platform. For makers of AI music, it’s a message: Spotify’s not banning AI completely, but it’s cracking down on spam, copycats, and low-effort tracks that push real artists out of the spotlight.
It’s part of a bigger worry inside the music world. Streaming services are getting swamped with AI-generated tracks—even stuff that mimics the styles of established artists and abuses royalty systems. Spotify’s not banning AI music outright, but it’s making it clear that fake, spammy content won’t make it into the playlists and mixes that drive discovery.
This latest move will probably heat up the debate—should streaming platforms ban bad AI music or just label and filter it? For now, Spotify’s strategy is to filter and label, not ban.
