Deezer reports that AI-generated tracks now make up about 44% of all new music uploaded daily, with nearly 75,000 AI tracks uploaded each day and over 2 million per month.

CEO Alexis Lanternier says Deezer’s AI-detection technology has cut AI-related fraud and payment dilution, and the company has offered licensing of this protection to industry peers since January to encourage adoption.

Deezer notes that AI-generated tracks account for roughly 1-3% of total streams, and detected fraudulent AI content is demonetized or removed from recommendations.

A major challenge cited is that watermarks can be removed from AI audio, and custom-model AI music without watermarking could worsen fraud and low-quality production as AI costs drop.

There is ongoing legal pushback against AI music tools alongside collaborations and licensing deals between major labels and AI music startups.

A test with producer YouTuber Lnkhey illustrates how listeners struggle to distinguish AI-generated music from human-made works.

Overall takeaway: AI-generated music is growing fast but faces mislabeling, fraud, and governance challenges that require clear labeling and policy action to protect creators and listeners.

Industry concern is rising in the U.S., with examples of AI-created country acts and a fully AI-generated group, The Velvet Sundown, gaining significant listenership in 2025.

The broader issue centers on artists’ visibility and rights as AI-generated vocals and acts draw attention in the U.S. music scene.

Other platforms like Coda Music are adopting AI-labeling and user-flagging to curb AI-generated content, signaling industry-wide countermeasures.

The trend shows AI music generation accelerating due to cheaper models (e.g., Lyria 3, Suno, Udio), with some services offering full-length AI tracks and watermarking to indicate AI origin.

Deezer has introduced an artist-centric remuneration cap to curb fraud, limiting streams per user and adjusting royalties when caps are exceeded.