Qobuz Says 60% of Streams on AI-Generated Music It Identifies Are Fraudulent
Qobuz says 60% of streams on tracks it identifies as AI-generated are classified as fraudulent by its anti-fraud systems and excluded from royalty payouts. The French streaming service is also making that identification visible to listeners, introducing labels on releases containing music it flags as AI-generated.
The company announced the rollout on September 24, with Music Business Worldwide reporting the development on September 28. The figures put streaming fraud, catalog management and artist compensation at the center of the story.
What the 60% figure means
According to Qobuz’s announcement, tracks identified as AI-generated account for just 0.38% of all streams on the platform. The 60% fraud figure applies specifically to streams within that identified group. It does not describe 60% of Qobuz’s overall listening activity or establish a fraud rate for AI-generated music across other services.
Qobuz says streams found to be fraudulent are excluded from reporting and royalty payments, and associated content is removed from its catalog. These are company-reported findings from its own detection systems.
Four million tracks affected by search cleanup
Qobuz also says it has removed from search more than a third of the albums it identified as AI-generated that had no listening activity. Those albums cover four million tracks.
That search cleanup is distinct from deleting fraudulent content from the catalog. As Music Business Worldwide notes, Qobuz’s updated policy allows de-indexed albums to remain findable through a direct title search. The change therefore concerns how listeners encounter those releases, rather than necessarily making them unavailable.
The release-level badge gives listeners another point of information when choosing what to play. Qobuz also keeps identified AI-generated content out of its editorial selections and algorithmic recommendations.
AI tools and human creativity
In its updated AI Charter, Qobuz supports AI tools used in an authentic human creative process, including mixing, mastering and composition assistance. Its stated objection is to industrially produced content without human creative intervention, intended to saturate catalogs or exploit royalty systems.
That distinction matters for electronic music producers. Using an AI-assisted production tool should not automatically be treated as equivalent to generating an entire recording without meaningful human creative involvement.
Detection remains imperfect. Qobuz acknowledges that no available solution can identify all AI content with certainty, particularly when human involvement is mixed into the process. Artists who believe their music has been incorrectly labeled can raise the issue through their distributor.
Why this matters for artists and listeners
The significance of the rollout is the connection between disclosure, discovery and payment. A label helps a listener make a choice; search and recommendation policies affect visibility; fraud controls determine which listening activity qualifies for royalties.
For artists, the practical questions are how accurately those systems identify abuse and how effectively they correct mistakes. For listeners, visible labeling makes a previously hidden platform judgment part of the listening experience.
Sources:
