Spotify removes 75M tracks, unveils AI disclosure standards

Spotify announced sweeping changes to its artificial intelligence music policies on Thursday, implementing new safeguards against deepfakes and spam while revealing it has removed over 75 million problematic tracks from its platform in the past year. The streaming giant’s three-pronged approach aims to combat AI-enabled fraud while supporting legitimate artistic use of the technology, marking its most comprehensive response yet to the surge of AI-generated content flooding music platforms.

New Standards Target AI Transparency and Spam

The company will adopt an upcoming industry standard for AI music disclosure developed by DDEX (Digital Data Exchange), allowing artists and labels to specify exactly how artificial intelligence was used in their creative process. Under this system, creators can indicate whether AI generated vocals, instrumentation, or handled post-production work, moving beyond a simple “AI or not AI” classification.techcrunch

“We know the use of AI is going to be a spectrum, with artists and producers incorporating AI in various parts of their creative workflow,” said Sam Duboff, Spotify’s Global Head of Marketing and Policy, during a press briefing. “This industry standard will allow for more accurate, nuanced disclosures.”techcrunch

Spotify will also roll out an enhanced music spam filter this fall targeting mass uploads, duplicate tracks, SEO manipulation, and artificially short songs designed to game royalty payments. The filter will identify problematic uploaders and remove their tracks from algorithmic recommendations rather than deleting them outright.billboard

Crackdown Follows AI Music Controversy

The policy overhaul comes in response to mounting controversy over AI-generated artists achieving viral success without proper disclosure. The Velvet Sundown, which amassed over 1 million monthly listeners on Spotify before confirming it was entirely AI-generated, exemplified the platform’s transparency problem. The band’s tracks appeared on popular playlists and users’ “Discover Weekly” recommendations despite Reddit users flagging suspicious elements like AI-generated profile photos and fake band member biographies.rollingstone

Spotify executives directly addressed persistent rumors that the platform deliberately promotes AI-generated content to reduce royalty payments. “Those rumors are categorically and absolutely false,” stated Sam Duboff during the policy briefing. “Spotify doesn’t generate any music. We don’t own any music. All the music on Spotify, 100% of it, is created, owned, uploaded by licensed third parties.” The company emphasized its commitment to supporting legitimate AI use while aggressively protecting against fraudulent schemes, such as the North Carolina musician who allegedly used AI to generate “hundreds of thousands” of songs to collect over $10 million in fraudulent streaming royalties.thenextweb