Spotify says it has pulled more than 50 million junk and AI-made songs out of its recommendations so far this year. The company announced it Tuesday. If your daily mixes have felt full of odd, generic tracks, the goal is for them to lean more toward human artists.
What changed
Earlier this year, Spotify looked at the tracks listeners skipped, ignored or said they did not want to hear. It found those tracks were consistently identified as slop or spam. Over the past six months, it has changed its recommendation system to weed them out and give music by human artists a boost.
Spotify also removed more than 75 million spammy tracks from the platform entirely. Those were for serious or repeated violations of its rules. Spotify says listener complaints about slop and spam in recommendations have dropped noticeably since the changes.
How Spotify spots it
The company says no single signal can catch all of it. It combines several, then judges whether a track is likely slop or spam.
- Listener engagement and feedback
- Upload patterns
- Track details, known as metadata
- Outside AI detection tools
Spotify also hunts for tricks such as mass uploads, duplicates, search manipulation, spammy playlist networks and artificially short tracks. It says it will keep updating its systems as spammers change tactics.
Why the money matters
Much of this is about fraud. In some schemes, a person generates hundreds or even thousands of AI songs, uploads them to streaming services, and uses bots to play them again and again to collect royalties. In 2024, Spotify set a 1,000-stream annual minimum for a track to earn recorded-music royalties. That made mass uploading less profitable.
Spotify has also started labeling AI-generated artist personas. It has verified more than 300,000 human artists, which account for over 99% of what listeners actually search for.
The wider fight
The music industry is still sorting out AI song tools like Suno and Udio. Universal Music Group and Sony Music Entertainment are still in court against Suno. They accuse it of using copyrighted recordings without permission to train its AI. Warner Music Group settled with Suno last year. In August, Suno said it plans audio watermarking and fingerprinting to trace its songs, and it has added download limits.
What is not known
The announcement, as reported, does not say how much AI music is still in recommendations or how Spotify defines slop. It also does not say whether listeners will see any visible change.
The takeawaySpotify is steering recommendations toward human artists, so skipping or flagging tracks you dislike helps shape what it treats as junk.