Exhibitor login
AI Insider 20 October 2026

Spotify removes 50 million AI tracks from recommendations

Spotify removes 50 million AI tracks from recommendations

In 2026, Spotify reduced the number of AI tracks and spam tracks in its recommendations by more than 50 million, while also removing 75 million tracks completely from the platform. This decision was made to provide listeners with a better experience, allowing them to hear more music from real artists in their playlists like Mix or Discover Weekly.

Looking back at their approach to AI spam, Spotify noted that users often skipped or clicked away from tracks recognized as AI spam. This led to several adjustments in the recommendation system, and the editorial team no longer includes such tracks in their playlists. However, detecting AI music is complex, as no single signal can identify AI spam on its own. Therefore, Spotify combines various signals, such as listening behavior and data from other AI detectors, to identify AI music.

Additionally, Spotify introduced labels like 'Verified by Spotify' for real artists and 'AI Persona' for AI-generated music tracks. This allows users to see which artists are actually behind a track. Users can also view the AI Credits in the app, showing which parts of a track were created with AI. By the end of the year, artists will have the opportunity to approve their new releases themselves through the Artist Profile Protection program.

While Spotify still allows AI music, it is becoming increasingly difficult for spammers to actually make money with such music. This is because tracks created with AI often do not reach the required 1,000 streams to be truly profitable. The recent European AI regulation, which will come into effect in August 2026, also emphasizes transparency for AI content. This indicates that it should be recognized what is created by a human and what is created by AI. Although Spotify does not produce the music itself, it is moving in the same direction toward recognizability for listeners.

Read the full article from AI Insider.