Deezer, a prominent music streaming platform, is taking a significant step in the fight against AI-generated music fraud by making its advanced AI detection tool available to other streaming services. Launched last year to automatically identify and remove fully AI-generated tracks from its recommendations, this initiative now aims to foster transparency across the music industry and ensure human artists receive due recognition and fair compensation.
The company revealed alarming statistics, reporting that 85% of streams from fully AI-generated tracks are deemed fraudulent. The scale of the issue is rapidly expanding, with Deezer now receiving approximately 60,000 AI tracks daily, bringing the total number of AI-detected songs on its platform to 13.4 million. This marks a substantial increase from June of last year, when AI-generated music constituted 18% of daily uploads, exceeding 20,000 tracks.
Deezer's AI music detection tool boasts a remarkable 99.8% accuracy rate, capable of identifying tracks generated by major models like Suno and Udio. Beyond merely excluding these tracks from algorithmic and editorial recommendations, the tool also demonetizes them, removing them from the royalty pool. This strategic move underscores Deezer's commitment to fairly compensating legitimate musicians and songwriters.
According to Deezer CEO Alexis Lanternier, the tool has garnered "great interest" from various companies, with several having already conducted "successful tests." Among the early adopters is Sacem, the influential French management company representing over 300,000 music creators and publishers, including renowned artists like David Guetta and DJ Snake. While specific pricing details and other interested parties remain undisclosed, a company spokesperson indicated that costs would vary based on the nature of the deal.
The music industry faces escalating concerns regarding AI companies utilizing copyrighted material for model training and sophisticated methods employed to manipulate streaming systems for fraudulent gains. A notable incident in 2024 saw a North Carolina musician charged by the Department of Justice (DOJ) for allegedly creating AI-generated songs and using bots to generate billions of streams, resulting in over $10 million in stolen royalties. The rise of AI bands, such as The Velvet Sundown, which have amassed millions of streams, further illustrates the evolving landscape.
Different streaming platforms are adopting varied approaches. Bandcamp recently implemented a complete ban on AI-generated music, while Spotify updated its policy to include labeling AI-used tracks, reducing spam, and explicitly prohibiting unauthorized voice clones. In contrast, major record labels like Universal Music Group and Warner Music Group have opted to embrace AI-generated music under licensing agreements with startups Suno and Udio, aiming to compensate artists whose work trains these AI models.
Deezer has consistently demonstrated a proactive stance on AI in music. In 2024, it became the first music streaming platform to sign the global statement on AI training, joining a host of notable creatives and actors. This latest move to share its robust detection tool is anticipated to set a crucial precedent, encouraging other music streaming platforms to implement similar measures to safeguard human artists and combat widespread fraud in the digital music ecosystem.
