Deezer's free AI music detector scans Spotify and Apple Music playlists as 44% of new uploads read as synthetic
Dezer is offering the tool to rivals' subscribers, sharpening a policy split over whether to label or remove AI tracks from the queue.
Dezer is offering the tool to rivals' subscribers, sharpening a policy split over whether to label or remove AI tracks from the queue.
The streaming era is about to get a lot more honest about what listeners are actually hearing. Deezer said on Thursday that 44% of the new music uploaded to its platform in recent weeks reads as AI-generated, and it is now offering a free tool that lets anyone scan playlists on Spotify, Apple Music, SoundCloud, YouTube Music, and roughly 16 other services to see how much of the catalog is synthetic.
The detector, announced via TechCrunch's coverage of the launch, supports 27 languages and works on tracks hosted by 20 of the most popular streaming platforms. A user pastes a playlist link, and the service returns a percentage breakdown of suspected AI tracks alongside a shareable report. The cross-platform reach is the point: Deezer is extending detection to users who pay its rivals, not only its own subscribers.
The 44% figure is what makes the tool more than a novelty. If that share holds across the industry, the question shifts from whether AI music is in your queue to how much of it is, and how to count it consistently. Deezer has framed the scale of the synthetic upload wave as the reason detection is now a trust issue, with copyrighted training data and potential streaming-fraud manipulation as the underlying risks.
The policy split among major platforms is sharper than the launch coverage suggested. Deezer has been one of the industry's most aggressive opponents of generative music, removing AI tracks from its algorithmic recommendations and editorial playlists rather than only labeling them. Apple Music and Spotify, by contrast, have stuck to a tagging approach that keeps AI music in playlists but marks it as synthetic, according to TechCrunch's reporting on the launch. Both postures are deliberate. Tagging preserves listener choice and avoids de-platforming artists who use AI as part of their process; removal protects editorial integrity and royalty pools but punishes a wider swath of synthetic output, including work that may have human creative input layered on top.
There is a precedent for the harder line. Bandcamp banned AI-generated music from its platform earlier, treating the question as binary rather than a matter of disclosure. Deezer has hinted it is weighing supplier-policy updates or content removal of its own, though it has not specified a timeline. The detection tool effectively gives the company, and listeners, a measurement layer to act on before any policy shift.
The legitimate criticism is in plain view regardless of which platform a user pays. Generative music models have been trained on copyrighted recordings without permission, which means a meaningful slice of the 44% may rest on material the original artists and labels never licensed for that use. And where detection lags, synthetic tracks can be deployed at scale to manipulate streaming payouts, a problem the industry has only partially mapped. The new scanner is one answer to the trust gap, but the underlying data practices and royalty safeguards are still unresolved.
What to watch next: whether Spotify or Apple Music respond by tightening their own detection or by changing their tagging language, and whether the 44% number holds up as the cross-platform scanner collects more samples. Deezer has effectively published a measurement the rest of the industry now has to react to.