Paris, 16:12 CET, January 30, 2026
- Deezer has licensed its AI music detection technology to France’s royalty agency Sacem in a commercial agreement.
- The streamer reported flagging over 13.4 million AI-generated tracks in 2025 and removed fake plays from its payout calculations.
- Deezer reports that AI-generated uploads now account for roughly 39% of daily submissions, driving efforts to boost adoption.
Deezer announced it has licensed its AI-generated music detection technology to France’s royalty agency Sacem. This marks a key commercial deal for the tool, which the streamer hopes will gain broader adoption across the music industry.
This shift comes amid generative AI’s rise, which simplifies producing large volumes of songs. At the same time, streaming services now grapple with a fresh type of fraud: bulk uploads designed to game recommendation algorithms and siphon royalties away from the rightful artists and songwriters.
Deezer reported it caught and removed as much as 85% of fake AI-generated music streams from its royalty pool in 2025, after flagging over 13.4 million AI tracks. The royalty pool, which CEO Alexis Lanternier says accounts for 70% of subscriber revenue, is the fund reserved for rights holders.
Uploads have surged quickly. Deezer now reports getting around 60,000 fully AI-created tracks daily, making up about 39% of its total uploads each day—up from just 10% in January last year.
The company explained that its detection tool analyzes audio signals for patterns tied to AI music generators like Suno and Udio, identifying subtle anomalies often missed by the human ear. Deezer revealed it trained the system on 94 million songs and secured two patents for the technology this year, 2024.
Lanternier noted that AI-created music is now “nearly indistinguishable” from tracks made by humans. He emphasized Deezer’s commitment to clearly labeling AI-generated songs and safeguarding rights holders. The platform also excludes AI music from its algorithmic recommendations, which typically boost play counts significantly.
Deezer said it “demonetizes” fraudulent streams by removing them from royalty calculations. The company is now offering its detection tech for licensing, following trials with Sacem. Newsroom Deezer
Deezer reported that fully AI-generated tracks make up just a small portion of its streams—up to 3%—yet fraud is rampant within that segment. Across its entire catalogue, streaming fraud represented 8% of all streams in 2025.
Sacem didn’t reply to Reuters’ request for comment. Deezer said it’s currently negotiating with other European collective societies on licensing and aims to connect with groups in Los Angeles during Grammy Week.
Other platforms are stepping up regulations on synthetic audio and impersonation, grappling with how to inform listeners about AI involvement. Spotify is introducing new policies and developing metadata standards, whereas Bandcamp has opted for a stricter approach, banning AI-generated content outright, The Verge reported.
Detection tools won’t end the bigger debate over authorship and copyright. Swedish royalty society Stim told Reuters that detection by itself can’t resolve issues around musical composition and rights, stressing that “copyright and technology can go hand in hand.” The group is advocating for mandatory licensing and full transparency on training data to tackle abuse right at its root.
There’s a real-world risk here: as AI generators get better, detectors might find themselves in a constant cat-and-mouse battle. Errors come with consequences — false positives could block legitimate releases, while misses mean royalties slipping through the cracks. Plus, adoption remains uncertain. A tool’s value depends entirely on how many big players actually use it regularly.
