Spotify said Oct. 20 that it has excluded more than 50 million tracks it classifies as spam or “slop” from recommendations in 2026, seeking to protect music discovery and artists’ potential royalties.
Those tracks have not necessarily been taken off the streaming service. Spotify separately said it has removed more than 75 million tracks from its platform for serious or repeated rule violations. It did not specify the period covered by that latest removal total.
In September 2025, Spotify reported removing more than 75 million spam tracks over the preceding 12 months. That count concerns tracks taken off the platform, while this year’s recommendation figure concerns tracks withheld from listeners’ suggested music.
To screen recommendations, Spotify uses track metadata, upload patterns, listener feedback and third-party tools that detect AI-generated content. The company said it changed its system over six months after studying tracks that listeners ignored, rejected or skipped. It acknowledged limits in AI detection and said it would keep adjusting the filters as tactics change.
Spotify has linked the problem to AI tools that make music easier to generate at scale. It says people seeking to exploit the service also use mass uploads, duplicates and artificially short tracks. In Spotify’s assessment, those practices can draw attention away from other artists and dilute the royalty pool.
The company also said it has verified more than 300,000 artists and applied its “AI Persona” label to thousands of profiles it believes represent AI-generated artist identities. It generally excludes those profiles from recommendations, though a listener who follows one may still encounter it there.
Spotify is testing Artist Profile Protection, which lets participating artists review releases that list them and control which appear on their profiles. It plans to make the tool available to all artists by the end of 2026.
