In brief: Qobuz wants its own detection system and clear labels. The solution is still being developed, and accuracy and implementation remain central questions.

Why build an in-house system?

Metadata supplied by distributors does not always show reliably how a recording was made. Its own analysis could help the platform make more consistent decisions about labels, recommendations and suspicious bulk uploads.

What distinguishes Qobuz’s position?

The service emphasises a highly curated, human-centred music experience, and its executives have been openly critical of fully automated content. A fair detector must still distinguish human work assisted by AI from entirely generated mass production.

Why does labeling matter?

Listeners can make an informed choice, while the platform can protect the quality of its recommendations. A label is useful only when it is understandable, verifiable and open to correction when the detector gets it wrong.

A seriously unserious ending

The great music exam

The detector asks the song: “Who wrote the chorus?” The song replies: “One human, three prompts and a cat that sat on the keyboard.” The system creates a new category.

Source: TechRadar — interview with Qobuz managing director Dan Mackta.