Spotify removes 50 million AI-generated tracks from recommendations, but not platform

Recommendation System Updates

Spotify announced that it has removed 50 million tracks categorized as "slop and spam" from its recommendation system in 2026. This action follows policy changes introduced over a year ago to address the increase of AI-generated audio on the platform. The company updated its recommendation algorithms to prioritize content from "authentic, human artists" and its editors no longer include identified slop or spam in curated playlists.

User Feedback and Detection Methods

The changes were prompted by user behavior, specifically high skip rates and negative feedback on certain tracks. Spotify observed a clear pattern of users rejecting content identified as slop or spam. The company uses various signals to detect such content, including metadata, user feedback, upload patterns, and third-party AI detection tools. Spotify acknowledges that AI detection tools have limitations and will face increasing challenges as generative AI technology advances.

Platform Removal Discrepancy

While 50 million tracks were removed from recommendations, the total number of "spammy tracks" removed from the platform due to policy violations remains at 75 million, a figure last updated in September 2025. This indicates that tracks removed from recommendations may still be available on the platform. A Spotify spokesperson confirmed the company continues to delete content that violates its rules but did not provide an updated total for platform removals.

Dual Approach to AI Content

Spotify's strategy appears to involve managing AI-generated content within its recommendation system while simultaneously integrating AI features for users. The company allows users to remix and create song covers using AI tools. This dual approach suggests Spotify is navigating the complexities of AI content, attempting to control unwanted AI-generated material while also exploring its creative applications.