Spotify Deletes 75 Million Songs in Massive Platform Purge
Spotify has pulled an estimated 75 million tracks from its platform in what is shaping up to be one of the largest content purges in streaming history. The Spotify AI music crackdown is not targeting artists who use artificial intelligence as part of their legitimate production process. Instead, it is aimed at something else entirely: mass-produced junk uploaded specifically to game the platform’s royalty system.
The distinction is critical. Spotify has been clear that this is not a ban on AI-assisted music. It is a direct response to what industry insiders are now calling “AI slop”—low-effort, low-quality tracks generated in bulk and dumped onto the platform with essentially zero creative intent behind them.
Why Spotify Is Cleaning House
According to Spotify senior director Sam Duboff, the core issue is scale. Text-to-song AI tools can now generate a finished track in seconds, and bad actors have been using that speed to flood the platform with thousands of nearly identical uploads. Roughly 100,000 new songs land on Spotify every single day, and industry estimates now suggest that the share of AI-generated uploads accounts for nearly half of all new music hitting streaming services.
The economics behind the scam are straightforward. Every stream draws from a shared royalty pool, and fraudsters have figured out how to exploit that pool by uploading massive batches of generic ambient noise, fake meditation tracks, and repetitive instrumental loops under made-up artist names. Bots then generate artificial plays on these tracks, often parked in background or focus playlists where listeners are not paying close attention to the credits.
The money siphoned off through this process comes directly out of the pool that would otherwise go to working musicians. For independent artists relying on streaming income to make rent, that dilution is not just an abstract industry problem; it manifests in actual missed royalty payments.
What Spotify Is Actually Targeting
Spotify has repeatedly stressed that AI tools themselves are not the enemy. Music made with AI assistance remains on the platform as long as creators hold the proper commercial rights and are using the technology as part of a legitimate creative process. The enforcement is squarely focused on abuse of the system, not the technology itself.
To that end, Spotify has bolstered its machine-learning detection systems to flag suspicious upload patterns and bot-driven streaming activity. The platform has also banned unauthorized AI voice cloning, meaning any track that mimics an established artist’s voice without permission is subject to immediate removal. Furthermore, Spotify has implemented new safeguards to prevent third parties from scraping its catalog to train AI models without a valid licensing agreement.
The broader industry is pushing for more transparency regarding these developments. Groups including the IFPI and the RIAA have been advocating for clearer labeling standards that would flag whether a track is AI-generated, AI-assisted, or fully human-made. The argument is that listeners deserve to know what they are streaming, and that transparency doubles as a layer of protection for artists whose work and likeness could otherwise get lost in the noise.
Spotify pulling 75 million tracks is a significant move, but it is unlikely to be a one-time fix. As AI music generation continues to become cheaper and faster, the pressure on streaming platforms to separate legitimate art from algorithm bait will only intensify.
