For every independent artist earning streaming royalties, 75 million uninvited AI tracks spent the past year quietly grabbing a share of that pool. Spotify confirmed it removed more than 75 million “spammy” AI-generated tracks from its catalogue over the past 12 months, according to the Los Angeles Times and Rolling Stone. The critical distinction: Spotify isn’t banning AI music. It’s going after fraud. The platform still welcomes AI-generated compositions — provided creators hold commercial rights, avoid unauthorized voice cloning, and show genuine creative intent. This surge in AI-generated content is part of a broader wave of AI infrastructure investment reshaping the digital landscape.
The Spam Machine Spotify Is Fighting
Mass uploads, fake metadata, and artificially short tracks turned royalty pools into an all-you-can-eat buffet for bots.
“AI slop” — a term Spotify executives themselves have used — describes tracks generated from simple prompts, uploaded in bulk, and engineered to game the algorithm. Think of it as the streaming equivalent of those SEO-stuffed content farms that polluted Google search results a decade ago, except now they produce ambient piano loops instead of keyword-stuffed blog posts. The problem isn’t new. In 2023, Spotify pulled tens of thousands of AI songs linked to the platform Boomy after detecting artificial streaming, according to Forbes. But generative tools got cheaper and faster, and the flood became a tsunami.
Here’s how Spotify’s updated policy fights back:
- Voice cloning ban: unauthorized AI replicas of recognizable artists are removed immediately, even if the uploader claims to be that artist
- Machine-learning spam filter: detects behavioral patterns — mass uploads, duplicate content, artificial streaming — then downranks or removes offending tracks
- DDEX metadata standard: labels and creators can tag whether a track is fully AI-generated or AI-assisted (think: AI vocals versus AI-assisted mixing), with Spotify surfacing this information to listeners
- Anti-scraping defenses: systems built to prevent generative-AI companies from bulk-harvesting the catalogue for model training
- Dedicated monitoring team: tracks evolving spam tactics and new attack vectors in real time
“Spammers [shouldn’t earn] royalties that should go to professional artists and songwriters,” Spotify policy leadership told Rolling Stone.
AI Music Is Still Welcome – With Strings Attached
Spotify draws a hard line between creative AI use and catalogue-stuffing, but not everyone agrees where that line falls.
Authorized AI remixes and covers? Fine. AI-assisted mixing and mastering? Also fine. Flooding the platform with thousands of thirty-second tracks tagged as “relaxing rain sounds”? That’s where you get purged. Spotify is actively working with labels to clear AI-created content for legitimate sale, according to Rolling Stone.
Not everyone is comfortable with the approach. Some AI-music advocates worry that aggressive filters could catch legitimate experimental work in the crossfire, or that labeling requirements stigmatize AI-assisted creativity relative to purely human music — a concern raised in industry commentary rather than Spotify’s official stance. Meanwhile, IFPI and RIAA are pushing voluntary labeling programs that are starting to resemble light-touch regulation.
The arms race won’t stop here. As detection sharpens, spam operations will adapt — blending minimal human edits into AI output to evade filters. The real question: will “human-made” become a premium playlist filter option before you even notice it’s missing?
