Spotify Addresses AI Slop and Spam
The company reported removing 50 million tracks, though questions remain about whether those files were completely wiped from the service entirely.
Over a year ago, Spotify announced policy updates to address the growing wave of AI-generated audio flooding its platform. In a recent blog post, the company provided an update on the initiative, though critics argue it still falls short of helping users avoid generative AI trash.
So far, Spotify states it has yanked 50 million tracks categorized as slop and spam from recommendations. Over recent months, it made several updates to the recommendation system to instead promote more tracks from authentic human artists, excluding identified spam from curated playlists.
These updates follow close observation of songs users skipped, ignored, or actively rejected. The company observed a clear pattern of listeners avoiding low-quality AI audio, leading to a marked decrease in concerns about such content in recommendations.
Spotify noted it must continuously update its detection systems as bad actors change tactics. Detection relies on metadata, user feedback, upload patterns, and third-party AI detection tools, despite the limitations of automated software as generation tech advances.
Despite these efforts, Spotify previously reported removing more than 75 million spammy tracks due to serious or repeated rule violations, matching figures cited in earlier company disclosures.
A company spokesperson indicated that deletions of rule-breaking content continue, even as exact removal statistics fluctuate.
Balancing AI Features and Moderation
Critics point out that Spotify appears to be trying to have it both ways by removing spam while simultaneously introducing generative AI tools that let users remix and cover songs.
Several suspected fake artist profiles continue to amass hundreds of thousands of monthly listeners, reportedly siphoning millions of dollars in revenue away from human musicians.
While aware that many users dislike generative AI audio, Spotify has primarily targeted recommendations rather than purging the platform entirely, complicating its stated commitment to putting human artistry first.
Unlike competitors such as Deezer, which plans to offer an explicit filter for generative AI audio, or Bandcamp and Tidal, which enforce stricter royalty and content rules, Spotify lacks a robust user-facing block option.
Company representatives have previously emphasized prioritizing harmful uses like spam and impersonation over filtering by creation method, though consumer advocates argue a toggle to block AI-tagged tracks should still be provided.
Meanwhile, artists are submitting thousands of AI credits daily to clarify technology usage. Spotify displays these disclosures in mobile views and applies virtual artist labels to keep them out of default recommendations.
The platform also relies on a streaming threshold requiring 1,000 streams before issuing royalties, arguing this successfully hinders bad actors from scaling low-quality AI audio generation for profit.
Additionally, over 300,000 human artists are now verified on the service, with Spotify expanding profile management tools to give musicians greater control over their catalogs.
