Deezer to purge fake and fraudulent AI tracks as uploads pass 50%
Developed over time first seen 3 weeks ago
Deezer has confirmed that fully AI-generated tracks made up more than half of all daily song uploads on its platform for the first time, peaking above 50% in June 2026, roughly 90,000 tracks. In response, the French streaming service has settled on tougher measures: it will remove any AI track suspected of being used in streaming fraud, along with AI tracks left unstreamed for six months or more, as part of an effort to protect artists' and songwriters' royalties amid the surge in synthetic content.
The move caps off an 18-month escalation in Deezer's AI-detection efforts, which began in January 2025 when AI-generated uploads stood at just 10% of the daily total; the company says it has since flagged over 13.4 million fully AI-generated tracks. Deezer has previously found that up to 85% of streams on such tracks were fraudulent, with manipulated streams already excluded from royalty payments, and detected songs barred from algorithmic recommendations and editorial playlists. Its detection tool only catches 100% AI-generated recordings from models such as Suno and Udio, not partially AI-assisted music, and CEO Alexis Lanternier called for industry-wide standards to build a "healthy and sustainable" music ecosystem.
- Fully AI tracks now over 50% of Deezer's daily uploads, ~90,000 songs.
- Deezer will delete AI tracks tied to fraud or unstreamed for 6+ months.
- Up to 85% of streams on AI tracks were previously found fraudulent.
New here? Start with this
Deezer is a French music streaming service that competes with the likes of Spotify and Apple Music, letting artists upload songs for listeners to stream in exchange for royalty payments. In recent years, AI tools such as Suno and Udio have made it possible to generate entire songs automatically, and these AI-made tracks have been flooding onto streaming platforms in growing numbers.
This matters because a share of these AI tracks are not just low-quality filler but are being used for streaming fraud, where automated or fake plays are generated to siphon royalty payments away from genuine artists and songwriters. Deezer has been developing technology to detect fully AI-generated recordings and has been tightening how it treats them, including excluding suspect streams from royalty calculations and keeping such tracks out of recommendations and playlists.
The wider issue reflects a challenge facing the whole music industry as AI-generation tools become more widespread and harder to distinguish from human-made music, raising questions about how streaming platforms can protect artists' income and listener trust without clear industry-wide rules yet in place.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Advocates of Deezer's crackdown argue it is a necessary defence of a royalty system built on trust: when up to 85% of streams on fully AI-generated tracks have been found fraudulent, every pound diverted to bot-driven plays is a pound taken from human musicians and songwriters who depend on streaming income to survive. They see removing fraud-linked and unstreamed AI tracks not as censorship of a technology but as basic platform hygiene, akin to removing spam, and argue that a functioning marketplace requires listeners and creators to trust that popularity reflects genuine human interest rather than manipulated metrics. Framing this as protecting a 'sustainable music ecosystem', supporters contend that without such standards, the entire economics of streaming risk collapsing under the weight of low-cost, mass-produced content designed purely to game the system.
The case against
Critics who are sympathetic to AI-assisted creators caution that a blanket purge risks conflating fraud with innovation, punishing legitimate musicians who use tools like Suno or Udio as part of a genuine creative process rather than a fraud scheme. They argue that detection systems, however well-intentioned, are imperfect and opaque, and that removing tracks simply for going unstreamed for six months penalises niche or slow-building work rather than targeting actual bad actors. There is also concern that a major platform unilaterally setting the rules for what counts as acceptable AI use, without broader industry or regulatory consensus, could entrench incumbents' interests and stifle a legitimate new mode of musical creation before society has properly worked out how it should be valued and credited.
More coverage
AI Entertainment Music Technology TV
Read the full article at the source →
Originally published by Billboard as “Deezer Says Daily Delivery of Fully AI Music Reaches More Than 50% for the First Time”.