Deezer opens AI detection tool to rival platforms to combat fraudulent music streams

Deezer Makes AI Detection Tool Available to Industry

Deezer announced it will make its AI detection tool commercially available to rival streaming platforms, marking a significant shift in how the music industry addresses AI-generated music 1

1

. The French music streaming service introduced the tool last year to automatically tag fully AI-generated tracks and remove them from algorithmic recommendations. Now, Deezer aims to extend this capability across the industry to combat fraudulent streams and promote music industry transparency 5

5

.

Deezer CEO Alexis Lanternier revealed there has been "great interest" in the tool, with several companies having already performed successful tests 1

1

. One notable adopter is Sacem, the French royalty agency representing over 300,000 music creators and publishers, including David Guetta and DJ Snake 3

3

. While Deezer hasn't disclosed pricing details, a spokesperson confirmed costs vary based on deal type 1

1

.

Alarming Scale of AI Music and Fraudulent Streams

The platform now receives 60,000 fully AI-generated tracks per day, totaling 13.4 million AI-detected songs in its database 1

1

. This represents roughly 39% of total daily uploads, a dramatic increase from 10% in January last year and 18% in June 3

3

. More concerning, Deezer reports that 85% of streams from fully AI-generated tracks are deemed fraudulent 1

1

.

The AI detection tool analyzes audio signals for patterns created by AI music generators such as Suno and Udio, identifying subtle anomalies inaudible to human ears 3

3

. Deezer claims the tool achieves 99.8% accuracy and can identify every AI-generated track from major generative models 1

1

. The company trained the system on 94 million songs and filed two patents for the technology in 2024 3

3

.

Protecting Human Artists' Royalties Through Demonetization

Beyond detection, Deezer takes active steps to protect human artists by excluding AI-generated tracks from recommendations and removing them from the royalty pool 1

1

. The platform demonetizes AI music entirely, ensuring payments flow solely to human musicians and songwriters based on verified streams 5

5

. "We know that the majority of AI-music is uploaded to Deezer with the purpose of committing fraud, and we continue to take action," Lanternier stated. "Every fraudulent stream that we detect is demonetized so that the royalties of human artists, songwriters and other rights owners are not affected" 2

2

.

Deezer successfully identified and removed up to 85% of fraudulent AI-generated music streams from its royalty pool in 2025 3

3

. The royalty pool comprises 70% of subscriber revenue, making this protection critical for artist compensation 3

3

.

Growing Concerns About Music Streaming Fraud

The music industry faces mounting concerns about copyright infringement and manipulation of streaming systems. In 2024, the U.S. Department of Justice charged a North Carolina musician with creating AI-generated songs and using bots to stream them billions of times, resulting in more than $10 million in stolen streaming royalties 1

1

. AI bands like The Velvet Sundown have gained millions of streams, blending into regular catalogs without disclosure 5

5

.

Rival streaming platforms are responding differently to this challenge. Bandcamp implemented a total ban on AI-generated music uploads 1

1

. Spotify began rolling out new policies to address AI music and impersonation, working on a new metadata standard for disclosing AI use while explicitly prohibiting unauthorized voice clones 2

2

.

Industry Debates Over Licensing and Training Data

Not everyone agrees that detection tools alone solve the problem. Swedish royalty society Stim told Reuters that "copyright and technology can go hand in hand," believing mandatory licensing and full transparency for training data would prevent fraud at its source 3

3

. Stim launched its own license in 2025 allowing AI companies to legally use copyrighted songs for training models.

Meanwhile, major record labels have taken a different approach. Universal Music Group and Warner Music Group resolved lawsuits with Suno and Udio last fall, striking licensing agreements that permit use of their music catalogs for model training with compensation flowing to artists and songwriters 1

1

.

Deezer continues expanding its reach, holding discussions with other European collective societies and planning engagement with organizations in Los Angeles during Grammy Week 3

3

. In 2024, Deezer became the first music streaming platform to sign the global statement on AI training, joining actors Kate McKinnon, Kevin Bacon, Kit Harington, and Rosie O'Donnell 1

1

.