Spotify says it removed 50 million slop and spam tracks from recommendations
The streaming service says the filtering effort uses listener behaviour, upload patterns and outside detection tools, while leaving room for disclosed AI-assisted music.
Spotify said it has removed more than 50 million “slop” and spam tracks from recommendations this year, following six months of changes intended to reduce the reach of low-quality and artificially generated music. The company said the measure applies to recommendations and editorial programming, rather than necessarily deleting every flagged track from its catalogue.
The company separately said it had removed more than 75 million spammy tracks from the platform for serious or repeated rule violations. Spotify said its editors do not program tracks identified as slop or spam, and that listener concerns about such material appearing in recommendations had declined markedly after the changes.
Its screening combines listener engagement and feedback with upload patterns, metadata and information from multiple third-party AI-detection tools. Songs that listeners skip, ignore or mark as unwanted consistently aligned with content identified as spam or slop, Spotify said. It also monitors mass uploads, duplicate releases, search-engine manipulation, spammy playlist networks and artificially short tracks.
Spotify acknowledged that no single indicator can reliably identify the material, and said AI-detection technology has limits that may become more pronounced as generative tools develop. Its platform rules prohibit deceptive impersonation and manipulated or synthetic media presented as authentic where it poses a risk of harm; enforcement can include content removal and account suspension or termination for repeated or serious breaches.
The recommendation crackdown is broader than Spotify’s existing labels for synthetic artist identities. As reported in August, Spotify had already said AI Persona profiles would be excluded by default from editorial and algorithmic recommendations. That badge concerns whether an artist’s public identity appears to be an AI-generated person, rather than how a recording itself was made.
Spotify has also introduced AI Credits, through which creators can disclose AI’s contribution to a song’s lyrics, voice models, instrumentals or other elements. The company supports the DDEX disclosure standard covering AI-generated vocals, instrumentation and post-production, and said creators were submitting tens of thousands of AI credits daily.
The distinction reflects Spotify’s stated approach: it says it does not treat the mere use of AI in music-making as a violation, while targeting tactics that can deceive listeners, manipulate engagement or divert royalties. More than 300,000 artists have been verified through its Verified by Spotify programme, more than 70% of them independent, according to the company.
Spotify said tracks showing a high probability of AI generation were 10 times more likely to remain below its royalty threshold. Recording royalties require at least 1,000 streams and a minimum number of unique listeners over the preceding 12 months.