Spotify purged 75M tracks in crackdown on AI-generated music spam

Spotify ↘ 2.02% has removed more than 75 million tracks flagged as spammy or abusive AI content from its platform over the past year, the company confirmed this week, as the streaming giant contends with an industrial-scale wave of artificial intelligence-generated music threatening to overwhelm its catalog and royalty system.

Sam Duboff, Spotify's Global Head of Artists, Marketing and Policy, said the company will not stand in the way of artists using AI as a creative tool but wants to ensure listeners know when they are hearing AI-generated content. The removal targets tracks designed to mislead listeners, impersonate artists through vocal deepfakes, or exploit the platform's recommendation and payment systems through mass uploads and duplicates.abc+2

A Flood of Uploads

More than 100,000 tracks are uploaded to Spotify every day, and the challenge of separating legitimate music from AI-generated spam has grown rapidly. Rival streaming service Deezer disclosed in April that 44 percent of all new uploads to its platform are now AI-generated, amounting to roughly 75,000 AI tracks per day — a 650 percent increase over 16 months. Industry analysts have pointed to these figures as indicative of a platform-wide problem across music streaming.pcmag+3

Despite the volume of AI-generated uploads, Duboff noted that actual listener engagement with such content remains minimal. "When we examine entirely AI-generated music, which raises concerns, we find that it constitutes well under 1 percent of total consumption," he said. "Listeners are not interested in it, and we observe very low engagement with such tracks."abc

New Policies and Transparency Measures

Spotify's crackdown involves a three-pronged approach announced in September 2025: an impersonation policy prohibiting unauthorized AI voice clones, a spam detection filter targeting mass uploads and artificially short tracks, and a new AI disclosure system built on the Digital Data Exchange (DDEX) metadata standard. Beginning in April 2026, Spotify started surfacing AI credits in its app, allowing artists to indicate where and how AI contributed to vocals, instrumentation, songwriting, or production.consequence+2

The company has also introduced a verified artist program with a green checkmark to help listeners distinguish authentic human creators. Duboff emphasized that Spotify is not banning AI music outright — rather, it is targeting spam, fraud, and impersonation while allowing transparent, properly disclosed AI-assisted music to remain on the platform.freeaimusictools+2

Protecting the Royalty Pool

The stakes are substantial. Spotify pays roughly $11 billion annually to artists through a stream-share royalty model. Under this system, any track generating streams draws from the same pool, meaning fraudulent or spammy AI uploads can divert payments from legitimate creators. By excluding flagged tracks from recommendations and removing the most egregious cases entirely, Spotify aims to ensure royalties flow to "verified, human, and authentic artists," as Duboff put it.afr+1