Crosswords Sudoku and Comics
Science

Deezer Will Delete AI-Generated Songs Used to Commit Streaming Fraud

The French music platform receives nearly 90,000 fully AI-generated tracks every day and found that 85% of AI music streams last year were fraudulent.

Als bedrijf dat zelf innovatie hoog in het vaandel draagt, verleent Piano's Maene graag zijn medewerking aan het IWT project "Music Following for Live Venues" van het jonge Gentse muziektechnologiebedrijf SampleSumo (i.s.m. MuseScore, iMinds en het Vlaams Proeftuin Platform).

In het kader van dit p
Als bedrijf dat zelf innovatie hoog in het vaande…      Music Streaming Smartphone    Thomas Bonte from Gent, Belgium / Wikimedia Commons (CC BY 2.0)
By Free News Press Editorial Team
Published July 22, 2026 at 1:15 AM PDT

More than half the songs uploaded to Deezer are generated by artificial intelligence, and the French streaming service says it is now removing them. The company announced it will delete AI-generated songs used for fraudulent streams, according to a report by CNET.

Deezer said it receives nearly 90,000 fully AI-generated tracks every day. The company identified more than 13.4 million AI tracks on its platform in the previous year alone. It also announced it will remove tracks that have not been streamed in six months or more.

The scale of the fraud is significant. Fully AI-generated songs make up only 1% to 3% of total streams on the platform, but 85% of those streams were fraudulent last year. Deezer said it does not pay royalties for streams it determines have been manipulated.

The way the scheme works is straightforward. Using tools such as Suno, Udio, and AIVA, fraudsters generate thousands of AI songs quickly, sometimes with simple text prompts to set genre, mood, instruments, and lyrics. They then bulk-upload the music and use streaming farms and bots to rack up millions of fake listens. The more streams accumulated, the larger the royalty payments collected from the streaming service. Earlier this year, a North Carolina man pleaded guilty to exactly this type of scheme, having gained more than $8 million in royalties.

The financial stakes for the broader music industry are growing. A recent study found that creators could lose nearly 25% of their revenues, upwards of $4 billion, by 2028 due to fraudulent AI music streams. Apple Music VP Oliver Schusser told The Hollywood Reporter in February that his division identified 2 billion fraudulent streams in 2025, preventing royalty payouts of $17 million.

Deezer is not acting alone. In late April, Spotify announced it would add verification badges to artist profiles so listeners could identify music made by real people. Last month, Tidal said it would label completely AI-generated music and remove fraudulent or deceptive tracks. Deezer itself began tagging albums containing AI content in June 2025 and has been excluding AI music from algorithmic and editorial recommendations since then.

The volume of AI music being created is enormous. Billboard reported in November that Suno's 2 million-plus users are generating 7 million AI songs every two weeks, more than the entire Spotify catalog.

Whether labeling and removal policies will be enough remains a question in the industry. CNET AI reporter Katelyn Chedraoui has written that AI labels and badges are not sufficient on their own, and that streaming giants will need to be extremely vigilant going forward.

授權方式及範圍:中華民國總統府│政府網站資料開放宣告
Authorization Method & Scope: 

Office of the President, Republic of China (Taiwan) | Government Website Open Information Announcement
授權方式及範圍:中華民國總統府│政府網站資料開放宣…      Music Streaming Smartphone    總統府 / Wikimedia Commons (CC BY 2.0)