Spotify Removes 50 Million AI Slop and Spam Tracks From Recommendations
The streaming giant purged millions of low-quality tracks over the past year, adjusting algorithms to prioritize human artists and reduce listener complaints.
Key points
- Spotify removed over 50 million AI-generated slop and spam tracks from its recommendation systems in the past year.
- The company deleted an additional 75 million spam tracks outright for serious rule violations.
- Spotify uses listener feedback, metadata, and third-party tools to identify non-human content, though it notes detection has limitations.
Spotify announced on Tuesday that it has excluded more than 50 million songs identified as AI-generated slop or spam from its user recommendations over the past year. According to the announcement, the streaming service also adjusted its recommendation algorithms to highlight music created by human artists, aiming to improve the listening experience and reduce exposure to low-quality content.
The removal of these 50 million tracks is distinct from the more than 75 million spam songs Spotify deleted outright last year for serious or repeated rule-breaking. The company did not specify how many of the excluded tracks were generated by artificial intelligence versus other forms of spam. Spotify noted that these excluded songs were consistently the ones listeners skipped, ignored, or explicitly rejected, indicating a clear mismatch between user preference and the recommended content.
Detection Methods and Industry Context
Spotify determines whether a track may be AI-generated slop by combining multiple signals, including listener engagement, upload patterns, metadata, and third-party AI detection tools. The company admitted that AI detection has limitations and will likely become more difficult as the underlying technology improves. Despite these challenges, Spotify stated that since implementing these changes, it has seen a marked decrease in listener concerns about this content appearing in recommendations.
Other music streaming platforms are adopting varied stances on AI-generated music. YouTube Music places labels on AI songs, while Deezer allows users to filter out AI slop. Bandcamp has banned AI music entirely, and TIDAL has cut off royalties for fully AI-generated tracks. In July, Deezer reported that more than half of the music uploaded to its service daily is now made with AI, highlighting the growing scale of the issue across the industry.
Artist Verification and Royalty Implications
Spotify has introduced a verified badge for human artists, which is now held by more than 300,000 acts. The service also labels thousands of AI-generated artists and keeps their music out of recommendations by default. Although Spotify has not mandated labels on all AI tracks, it launched an “AI Persona” badge for artists that use an AI identity and do not represent a real person. Music from these AI personas is not recommended by the platform.
The financial impact on AI-generated content is significant. Spotify stated that tracks likely to be AI-generated are ten times more likely to earn fewer than 1,000 streams a year. This threshold is the minimum required for a track to earn royalties on the service. The company is also partnering with Universal Music Group and the independent music publisher Believe to launch a tool for creating AI music remixes, signaling a continued, albeit controlled, role for AI in music production.
What should you do?
Security and technology professionals monitoring content integrity should observe how platforms like Spotify balance automated detection with manual curation. As AI generation capabilities improve, detection methods relying on metadata and upload patterns may become less reliable. Organizations should prepare for evolving standards in content labeling and verification to maintain trust with their user bases.
What to do and how to stay safe: Spotify
- Monitor official announcements from streaming platforms for updates on content moderation policies and AI detection capabilities.
- Review your own content distribution strategies to ensure compliance with evolving platform rules regarding AI-generated media.
- Verify the authenticity of artist profiles and tracks by checking for verified badges or official labels provided by the service.
- Stay informed about industry-wide trends, such as royalty restrictions for AI-generated content, which may affect long-term viability.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How many AI-generated tracks did Spotify remove?
Spotify removed more than 50 million tracks identified as AI-generated slop or spam from its recommendations over the past year.
Does Spotify delete all AI-generated music?
No, Spotify does not delete all AI-generated music; it excludes certain tracks from recommendations and labels AI personas, but the songs remain on the service.
How does Spotify detect AI-generated slop?
Spotify uses a combination of listener engagement, upload patterns, metadata, and third-party AI detection tools to identify potential AI-generated slop.
