Qobuz is putting a visible label on releases it identifies as AI-generated, while also tightening where that music can be discovered on its service. The French streaming platform now displays the message "This release includes AI-generated tracks" at release level, rather than simply leaving listeners to infer how a track was made.
The move is paired with changes to search, recommendations and royalty handling. [Music Business Worldwide](https://www.musicbusinessworldwide.com/qobuz-starts-tagging-ai-generated-music-in-its-app-and-says-60-of-streams-on-ai-tracks-are-fraudulent/) reports that Qobuz has removed more than a third of the AI-identified albums with no listening activity from its search engine. Qobuz says those albums cover four million tracks, although they can still be found through a direct title search.
The company's updated AI Charter also keeps identified AI content out of Qobuz's human-curated editorial areas, including its new-release and playlist surfaces. It is excluded from algorithmic recommendations too. The charter names Our Selections, New Releases, Qobuzissimes and Albums of the Week among the editorial surfaces where such content will not be promoted. That makes the tag more than a disclosure device: it sits alongside decisions about what the platform will actively surface.
Qobuz says its detection system was developed in-house. [Music Ally](https://musically.com/2026/09/24/ai-roundup-landr-talks-fraud-while-qobuz-tags-genai-tracks/) reported that the service describes its approach as similar to fellow French DSP Deezer's. The platform says AI-generated music accounts for 0.38% of all streams on Qobuz, a figure that should be read as the company's own assessment rather than an independent measurement.
Its other headline number carries the same limit. Qobuz says 60% of streams on tracks it has identified as AI-generated are currently deemed fraudulent by its anti-fraud systems. Those streams are excluded from royalty payments and reporting, according to the company. Neither report independently audits the detection method or the fraud assessment.
That uncertainty is acknowledged in Qobuz's own charter. It says no reliable solution currently identifies all AI content with certainty, particularly where there has been partial human involvement. Artists or rightsholders who believe a release has been labelled incorrectly are directed to their distributor, while listeners can report music they believe has not been tagged.
The platform is also working on the DDEX field intended to identify AI-generated content when it is declared by rightsholders. Qobuz says the rollout is under way across labels, distributors and aggregators. That route is different from its in-house detection, but both are now part of the same policy framework. The charter was first published in February and has now been updated for September, according to Music Business Worldwide.
The practical picture is deliberately narrow. Qobuz has not said it is removing all AI-made music, and a badge does not settle every question around a release's creation. The sources do not provide an outside test of its detector, nor do they establish how often a badge is disputed. Qobuz says it will continue revising the charter as technology and music-industry regulation change.
What it has done is make its own labels visible to listeners, reduce the reach of identified content in its discovery systems, and set out how disputed tags can be challenged. The policy joins its detection, rightsholder declarations and anti-fraud measures in one framework, while leaving the limits of each measure plainly stated.
Read more on EDM Global News.