Qobuz Starts Tagging AI-Generated Releases — and Says 60% of Their Streams Are Fraudulent

Qobuz has started labelling releases it believes contain AI-generated music. From 24 September 2026, the French hi-res streaming and download service began surfacing a tag inside its app on releases flagged by a detection system it says was designed and developed in-house.

The wording of the label, as reported by Music Ally, is straightforward: "This release includes AI-generated tracks", sitting alongside a link where a listener can read more about what the tag means. There is no grading, no confidence score shown to the user — just a flag on the release page.

Alongside the rollout, Qobuz published a set of numbers about what its systems are finding. Those figures, rather than the tag itself, are the more revealing part of the announcement.

The numbers Qobuz put on the table

According to Qobuz, tracks identified as AI-generated account for 0.38% of total streams on the platform. That is a small share of listening, and the company is clearly keen to frame it that way. But the second figure is the one that has travelled furthest: Qobuz says 60% of the streams those AI-identified tracks do receive are currently deemed fraudulent by its anti-fraud systems, and are therefore excluded both from reporting and from royalty payments.

In other words, on Qobuz's own account, the majority of listening activity attached to AI-generated material on the service is not listening at all. It is the signature of streaming fraud — the bot-farm economics that have shadowed the streaming business for years, now with a limitless supply of cheap content to attach itself to.

Qobuz also says it has removed from search visibility more than a third of the albums it identified as AI-generated that had recorded no listening activity whatsoever, a sweep that covers roughly four million tracks. Those releases are hidden from search rather than presented as a wholesale purge of synthetic music from the catalogue; the company's stated position is that listeners should be able to identify AI-generated work, not that it disappears.

An update to the AI Charter

The tag is presented as a step under the AI Charter that Qobuz first published in February 2026, a document the company says it has now updated to reflect these developments. The charter has been Qobuz's framing device for the whole issue: transparency for listeners, protection of royalty pools, and human editorial as the counterweight to algorithmic sludge.

That last point matters to how Qobuz positions itself. The service says its discovery — selections, playlists, featured content — is driven by human editorial teams, and that this human-curated approach keeps AI-generated content out of what it puts in front of listeners. Where a platform leans on algorithmic recommendation, mass-uploaded synthetic tracks have an obvious route to an audience. Where a human picks the playlist, they do not.

Deputy CEO Georges Fornay put the underlying argument plainly in comments carried by StereoNET: "Everyone should be able to know what they are listening to and make informed choices." The press release also has Fornay framing the problem as having moved on from detection to communication — that the challenge is no longer just spotting AI-generated music but letting listeners identify it clearly.

What Qobuz has not said

Qobuz has not published how the detection system works, how accurate it is, or what its false-positive rate might be. That is a meaningful gap: a tag applied to a release is a claim about that release, and there is currently no public way to assess how often the claim is right.

To its credit, the company is not pretending otherwise. StereoNET reports that Qobuz acknowledges detection remains imperfect — particularly for hybrid works where human and machine contributions are tangled together — and that industry-wide standards will eventually be needed. A track with AI-assisted mastering, an AI-generated backing pad, or a synthetic vocal double sits in a grey zone that a binary release-level flag cannot really describe.

Qobuz joins a pattern

Music Ally notes that the approach mirrors that of Deezer, the other French streaming service, which has been the most vocal operator on this subject. StereoNET's roundup of the wider landscape gives a sense of scale: Deezer said in April 2026 it was receiving around 75,000 fully AI-generated tracks a day, amounting to 44% of new uploads. TIDAL strips royalties from wholly AI-generated music and carries listener-facing labels. Spotify introduced AI persona badges for synthetic artists in August 2026.

Set against Deezer's 44% of daily uploads, Qobuz's 0.38% of streams is a reminder that upload volume and listening volume are entirely different quantities. The flood is real; the audience for it, at least on a subscription service built around hi-res catalogue and editorial, appears not to be. And if 60% of even that sliver is fraudulent, the genuine human appetite for machine-generated tracks on Qobuz is smaller still.

What the labelling wave has in common across Deezer, TIDAL, Spotify and now Qobuz is an implicit concession: the catalogue can no longer be assumed to be human by default. Disclosure is becoming a metadata field like any other — and the argument has shifted from whether listeners should be told to how reliably anyone can tell.

Further coverage of the announcement is available from Hifi Pig and Digital Music News.

This article was written by an AI system from Into The HiFi AI Desk, generated from the following sources, with no human editing pass before publication: Qobuz press release via Shore Fire Media — "Qobuz Rolls Out a Tag to Flag AI-Generated Music" (24 Sept 2026), Music Ally — "AI roundup: Landr talks fraud while Qobuz tags GenAI tracks" (24 Sept 2026), Hifi Pig — "Qobuz Rolls Out AI-Generated Music Tag" (24 Sept 2026), StereoNET — "Qobuz Starts Flagging AI-Generated Music For Listeners" (25 Sept 2026), Digital Music News (24 Sept 2026).