Spotify pulled 50 million AI slop and spam tracks from its recommendations

Listeners on Spotify will discover fewer automatically generated tracks and more human-made music in their recommendations.

Over the past year, Spotify has removed over 50 million tracks identified as AI-generated spam or "slop" from its user recommendations. The music streaming giant announced Tuesday that these actions, coupled with system adjustments, aim to elevate music created by human artists. Spotify's editorial team maintains a strict standard, excluding AI-generated spam from programming. Since implementing these measures, Spotify reports a significant reduction in listener complaints regarding such content in recommendations. While not mandating labels for all AI tracks, Spotify has introduced an "AI Persona" badge for artists using AI identities separate from real individuals and will not recommend music from these personas. This move follows similar actions by other platforms, with YouTube Music labeling AI songs, Deezer offering AI music filtering, and Bandcamp banning AI music altogether. Spotify employs a multi-faceted approach to identify potential AI-generated spam, including listener data, upload patterns, metadata analysis, and third-party AI detection tools. Last year, Spotify removed over 75 million spam songs and is collaborating with Universal Music Group and Believe on an AI music remix tool.

Spotify's significant removal of AI-generated content indicates a growing challenge for streaming platforms to curate authentic audio experiences amidst rapid AI advancements. By prioritizing human-created music and implementing a nuanced labeling system for AI personas, Spotify is attempting to balance technological innovation with artist and listener expectations. This strategy reflects a broader industry trend of grappling with AI's implications for music creation, copyright, and fair compensation.

The varied approaches across platforms, from Bandcamp's outright ban to Spotify's selective filtering and YouTube's labeling, suggest an ongoing market exploration for AI music governance. Spotify's reliance on a combination of user feedback, metadata, and detection tools highlights the difficulty in definitively identifying and managing AI-generated content, setting the stage for potential future disputes and evolving industry standards.