Deezer reports that 97% of listeners struggle to distinguish AI-generated music from human-made tracks, despite a surge in AI uploads.
Deezer, a French music streaming service, has revealed that AI-generated music now constitutes 44% of its new daily uploads, amounting to about 75,000 new compositions each day, or over 2 million monthly. However, these AI tracks attract minimal listener engagement, representing only 1-3% of total streams.
The disparity between the number of AI tracks uploaded and their actual listens is largely due to identified fraudulent plays. Deezer's technology has ascertained that approximately 85% of plays for these AI tracks are not genuine, leading to their exclusion from the platform's royalty payment system.
In January 2025, Deezer introduced the Deezer AI Detection tool, becoming the first music service to label AI-generated content. This tool boasts a remarkable accuracy rate of 99.8% and has identified over 13.4 million AI tracks in its catalog over the past year. Deezer's CEO, Alexis Lanturnier, emphasized the growing significance of AI music, calling for industry-wide collaboration to safeguard artists' rights and promote transparency.
In a recent study involving 9,000 participants across eight countries, Deezer found that an overwhelming 97% of listeners were unable to differentiate between human-made music and that created by AI. Furthermore, 80% of respondents expressed support for mandatory labeling of AI-generated content, while 73% indicated a desire to know if a service recommends entirely AI-generated music. Additionally, 52% of listeners believe such tracks should be excluded from main music charts.
Concerns about AI impersonating artists have also surfaced, highlighted by the case of jazz pianist Jason Moran, who discovered an album released under his name on Spotify without his consent. After investigation, it was revealed that AI bots were using his identity to release music. Spotify has acknowledged this issue, reporting the removal of over 75 million "spam tracks" in the past year and implementing stricter regulations against impersonation.
In February 2026, Google launched Lyria 3, an AI model designed for music generation that allows users to create audio from text prompts, images, or videos, further complicating the landscape for music authenticity.
