Deezer–Ipsos survey finds low recognition of AI-made music

At the same time, the survey results show continued demand for labeling. Seventy‑three percent of participants want clear tags for AI‑generated music on streaming services, 45% support filters that let them exclude such tracks from recommendations, and another 40% say they would prefer to skip them.

Seventy‑one percent also admitted that identifying a track’s origin remained difficult even after repeated listening. These figures set an agenda for platforms: transparent content identification and clearer explanations of how recommendations are built.

Streaming services are trying to balance moderation with monetization. By Deezer’s estimate, tens of thousands of AI‑generated tracks are uploaded every day, accounting for roughly a third of all new submissions. The company is rolling out tags for such content, excluding it from editorial playlists, and removing artificial streams from royalty calculations. Additional measures under discussion include more precise rights attribution and separated payouts for artists, producers, and creators of generative models.

The market continues to flag abuse and verification lapses. From time to time, projects created entirely by algorithms climb into charts, with their origin becoming public only later. Rightsholders point to risks from unauthorized use of source recordings in model training. Policy debates now cover licensing options, the scope of permitted copying, and technical controls for tracing training data.

In practical terms, music platforms face a need to revisit labeling, adjust recommendation engines, and refine royalty schemes. In the near term, the industry will keep looking for a balance between accessible technology, legal clarity, and the listening experience audiences are used to.