28% of Uploads Are AI-Generated: What Deezer's Data Tells Us

Deezer's head of public affairs recently disclosed a staggering figure: approximately 50,000 AI-generated tracks are uploaded to their platform every single day. That's roughly 18 million tracks per year from AI tools alone. Yet AI music captures only about 0.5% of total streams.

The gap between volume and engagement tells us two things. First, the market for AI-assisted music creation is real and growing exponentially. The tools are accessible, the output quality is improving rapidly, and creators are using them at massive scale. Second, the vast majority of AI-generated content lacks the differentiation, discoverability, and trust signals that drive meaningful audience engagement.

For creators who invest genuine human creativity — writing lyrics, shaping arrangements, making editorial choices about production — this gap is actually an opportunity. The flood of undifferentiated AI output makes provenance documentation more valuable, not less. When 50,000 tracks per day have no documented authorship chain, the ones that do stand out.

The platform response is telling. Deezer has been among the most transparent major platforms about AI music volumes, and they've signaled interest in provenance metadata as a quality signal. If platforms begin weighting content with verified human authorship in their recommendation algorithms — a direction that C2PA adoption and EU AI Act compliance are pushing toward — documented provenance becomes a competitive advantage.

The industry is moving from a world where all music was assumed to be human-made to one where that assumption no longer holds. In this new reality, proof of human contribution isn't a legal formality — it's a market differentiator.