Spotify is doing more to address AI 'slop' on its platform (www.engadget.com)

Spotify announced a set of policy changes to curb AI-driven spam and improve transparency around AI use in music. Working with metadata standard body DDEX, the streamer will introduce granular AI disclosures in credits (specifying AI-generated vocals, instrumentation, or post‑production rather than a binary “AI/no AI” flag), roll out a new impersonation policy to protect artists from unauthorized voice cloning, and launch a spam filter this fall to detect and tag mass uploads, duplicates, SEO hacks, artificially short tracks and other “slop.” Spotify says it’s already removed more than 75 million spammy tracks in the past year and will stop recommending identified spam to users. For the AI/ML community, this signals a push toward standardized provenance metadata and operational detection systems at scale. Granular disclosure metadata helps downstream systems (recommendation, curation, royalties) make finer distinctions, while the spam filter implies production-scale classifiers combining content analysis, behavioral heuristics and metadata validation to surface misuse. The move acknowledges both creative uses of generative models and the risk that low-cost AI tools enable bad actors and content farms to dilute royalties and listener trust. Spotify stops short of banning fully AI-made projects (e.g., The Velvet Sundown remains), instead prioritizing transparency, artist recourse and cross‑industry standards to manage AI-generated content.