Music streaming service Deezer starts tagging AI-generated music to tackle streaming fraud
Music streaming service Deezer has announced the launch of an AI tagging system to clearly label albums containing fully AI-generated tracks, including those made using models such as Suno and Udio. This update aims to deliver greater transparency for listeners concerned about content origin.
Alongside this move, Deezer has revealed that approximately 18 percent of daily uploads — over 20,000 tracks — are entirely produced by artificial intelligence. However, only about 0.5 percent of overall streams on the platform come from fully AI-generated music, and the majority of these AI-uploaded tracks are never played.
In response to fraud concerns, Deezer reports that as many as 70 percent of streams from fully AI-generated tracks are artificially manipulated, leading to those streams being excluded from royalty payments. The platform also notes that the high volume of unplayed AI music contributes to catalog dilution and is often connected to fraudulent activities.
To further protect artist remuneration and enhance listener experience, Deezer is now excluding fully AI-generated tracks from algorithmic and editorial recommendations. These measures reflect Deezer’s evolving strategy for handling the impacts of generative artificial intelligence on the music industry.

Comments
When AI create more problems that it's supposed to resolve... The Deezer tool actually detects fully AI-generated music, and may be kind of bad at recognizing music generated by AI at 99%, or mixed music from different AI. it's also the same problem with all CSAM detectors, that can be easily fooled and even last AI-based ones don't show much progress. As for humans, AI is sometime convincing enough most won't differentiate some talent and hard work from a good prompt, and so, which one should earn more money than the other?