Last year, Deezer introduced an AI detection tool that automatically highlights fully AI-generated music for listeners and removes it from algorithmic and editorial suggestions.
See also: Can AI make classical music like a composer?
The company announced Thursday that it is now making the tool available to other streaming platforms in an effort to address the rise of AI and fraudulent streams, as well as promote transparency in the music industry and ensure that human artists still receive the recognition they deserve.
Alongside this move, Deezer said that 85% of streams of fully AI-generated tracks are considered fraudulent. Notably, the service now receives 60,000 AI-generated tracks per day, totaling 13.4 million AI-detected songs. In contrast, in June last year, fully AI-generated music made up 18% of daily uploads, exceeding 20,000 tracks.
Deezer claims that its AI music detection tool can identify any AI-generated track from leading genetic models like Suno and Udio. In addition to excluding AI-generated tracks from recommendations, Deezer's tool financially isolates them and excludes them from the royalty fund, as the company aims to fairly compensate musicians and composers.
The tool's accuracy is 99.8%, a company spokesperson told TechCrunch.
Deezer CEO Alexis Lanterniersays there is “a lot of interest” in the tool, and several companies have “already carried out successful trials.” One of those companies is Sacem, the French management company that represents over 300,000 music creators and publishers, including David Guetta and DJ Snake.
The company did not provide pricing information or reveal which additional companies are interested in adopting the tool. A spokesperson said the cost varies depending on the type of deal.
See also: Create Apple Music Playlists with ChatGPT!
There is growing concern about AI companies using copyrighted material to train their models, as well as the methods used to manipulate streaming systems and commit fraud.
An example of music streaming fraud occurred in 2024, when a North Carolina musician was charged by the Department of Justice (DOJ) with creating AI-generated songs and using bots to play them billions of times, resulting in more than $10 million in stolen streaming rights. Additionally, AI bands like The Velvet Sundown have garnered millions of streams.
Bandcamp banned AI-generated music entirely, while Spotify has updated its policy to address the rise of AI-generated tracks, clarifying when AI is used in music production, reducing spam, and explicitly stating that unauthorized vocal clones are prohibited on the platform.
In contrast, major record labels have settled lawsuits with Suno and Udio, appearing to embrace AI-generated music. Last fall, Universal Music Group and Warner Music Group struck deals with these AI startups to license their music catalogs, ensuring that artists and composers are compensated when their work is used to train AI models.
In recent years, Deezer has taken significant steps to address concerns about AI-generated music. In 2024, it became the first music streaming platform to sign the International Declaration on AI Education, joining actors like Kate McKinnon, Kevin Bacon, Kit Harington, Rosie O'Donnell and other well-known creators.
See also: Pirate library “downloaded” 86 million Spotify songs
We hope that Deezer's latest decision to sell its detection tool will set a precedent for other music streaming platforms to take similar steps to defend human artists and combat fraud.
