Deezer says its music streaming platform receives nearly 75,000 AI-generated song posts every day, accounting for about 44 percent of all daily uploads, as previously reported by TechCrunch. Despite the rise in “illegal” uploads, Deezer said:
issued 29 minutes ago above April 22, 2026, 2:00 p.m.
by web desk
Deezer says its music streaming platform receives nearly 75,000 AI-generated song posts every day, accounting for about 44 percent of all daily uploads, as previously reported by TechCrunch. Deezer said consumption of AI songs now accounts for about 1% to 3% of total streams, as the platform continues to remove AI-generated music from its recommendation algorithms, despite a rise in “illegal” uploads. Deezer positions the tool as setting an “industry standard,” adding that the platform is currently the “only” music streaming service that tags AI-generated tracks. The service also demonetized AI-generated songs and stopped storing high-resolution versions of them. “As AI-generated music is no longer a marginal phenomenon and the number of streams continues to increase daily, we hope the entire music ecosystem will join us in taking action to protect artists’ rights and promote transparency for fans,” Deezer CEO Alexis Lanternier wrote in a blog post. As AI-powered song creation tools like Suno and Udio grow in popularity, music streaming services are dealing with the flood of AI tracks in a variety of ways. Spotify announces new policy to crack down on AI-generated music, while Apple Music is asking artists and record companies to label songs made with AI. Bandcamp banned AI music outright, and Qobuz similarly started automatically detecting and labeling AI music. According to Deezer, since launching the tool in January 2025, the number of AI-generated tracks uploaded daily has jumped from 10,000 to 75,000. Deezer has also started allowing other companies to license its AI song detection tools. As Deezer noted, the tool can identify songs created using Udio and Suno, and “could essentially add detection capabilities to other similar tools, as long as they have access to relevant data examples.” The company is also working on ways for its music detection tool to identify songs without the need for a training dataset.
