AI NewsGo-to-marketAnnouncement

Spotify pulled 50 million AI slop tracks from its recommendations

Spotify published a one-year review of its AI protection policy that reports more than 50 million tracks pulled from recommendations as slop or spam, 75 million taken off the platform entirely, and an AI-generated track 10 times more likely to fall below the 1,000-stream royalty threshold.

AI News

Editorial2 min read

LinkedInX
Spotify Newsroom header for the Prioritizing Human Artistry announcement, showing a stylised wave of audio bars

Image: Spotify

Why it mattersThese numbers tell a software team what the backlog looks like at a platform that already has AI detection, mandatory disclosure and verified-artist labels in production, and they say that even those tools do not stop the slop at the upload step.

Spotify published a one-year review of its AI protection work this week, under the title "Putting Human Artistry First on Spotify." The announcement, posted to Spotify's own newsroom, reports more than 50 million tracks removed from recommendations as AI slop or spam over the past year, more than 75 million taken off the platform entirely, and a set of new labels and policies that go with them.

Emma Roth at The Verge reports the same figures and adds that Spotify changed its recommendation system to highlight music by human artists. The Verge quotes Spotify saying "our editors hold the same standard and don't program content identified as slop or spam."

What Spotify counts and how

The newsroom post lists five pieces of policy in production: an AI Credits field that artists use to disclose which roles and instruments were AI-generated, visible in Song Credits on mobile and launched in April; a Verified by Spotify badge that covers 99 percent of listener searches; an AI Persona label that identifies AI-generated artist identities and excludes them from default recommendations unless the listener follows them; artist review of releases before they appear on an artist profile, now expanding to all artists by year end; and the 1,000-stream royalty threshold that has been live since 2024.

The detection method is a combination of listener engagement signals, feedback, upload patterns, metadata and third-party AI detection tools. Spotify says tens of thousands of AI Credits are submitted every day, and that more than 300,000 artists are now verified, 70 percent of them independent.

The number that stands out

The most useful figure for a team outside music is a comparison inside the Spotify data. The newsroom post says an AI-generated track is 10 times more likely than a human one to fall below the 1,000-stream royalty threshold. The threshold is Spotify's existing spam filter, and the ratio describes who is uploading what. Last year Spotify said it removed more than 75 million spam tracks from the platform in total.

Spotify has detection, mandatory AI disclosure, verified-artist labels and a royalty threshold that already functions as a spam filter. Even with those in production, 50 million tracks a year still have to be pulled from recommendations by hand. The newsroom post names what Spotify built; it also tells any platform that accepts AI content at upload time what the review workload looks like a year in.

Source

One Year In: Putting Human Artistry First on Spotify, Spotify Newsroom. Reporting by Emma Roth, The Verge.

SourceSpotify

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project