How this Parisian music streaming service is fighting AI fraud

Music streaming service Deezer said Friday that it will start flagging albums with AI-generated songs, part of its fight against streaming fraudsters.

Deezer, based in Paris, is grappling with a surge in music on its platform created using artificial intelligence tools it says are being wielded to earn royalties fraudulently.

The app will display an on-screen label warning about “AI-generated content” and notify listeners that some tracks on an album were created with song generators.

Deezer is a small player in music streaming, which is dominated by Spotify, Amazon and Apple, but the company said AI-generated music is an “industry-wide issue.” It’s committed to “safeguarding the rights of artists and songwriters at a time where copyright law is being put into question in favor of training AI models,” CEO Alexis Lanternier said in a press release.

Deezer’s move underscores the disruption caused by generative AI systems, which are trained on the contents of the internet including text, images and audio available online. AI companies are facing a slew of lawsuits challenging their practice of scraping the web for such training data without paying for it.

According to an AI song detection tool that Deezer rolled out this year, 18% of songs uploaded to its platform each day, or about 20,000 tracks, are now completely AI generated. Just three months earlier, that number was 10%, Lanternier said in a recent interview.

AI has many benefits but it also “creates a lot of questions” for the music industry, Lanternier told the Associated Press. Using AI to make music is fine as long as there’s an artist behind it but the problem arises when anyone, or even a bot, can use it to make music, he said.

Music fraudsters “create tons of songs. They upload, they try to get on playlists or recommendations, and as a result they gather royalties,” he said.

Musicians can’t upload music directly to Deezer or rival platforms like Spotify or Apple Music. Music labels or digital distribution platforms can do it for artists they have contracts with, while anyone else can use a “self service” distribution company.

Fully AI-generated music still accounts for only about 0.5% of total streams on Deezer. But the company said it’s “evident” that fraud is “the primary purpose” for these songs because it suspects that as many as seven in 10 listens of an AI song are done by streaming “farms” or bots, instead of humans.

Any AI songs used for “stream manipulation” will be cut off from royalty payments, Deezer said.

AI has been a hot topic in the music industry, with debates swirling around its creative possibilities as well as concerns about its legality.

Two of the most popular AI song generators, Suno and Udio, are being sued by record companies for copyright infringement, and face allegations they exploited recorded works of artists from Chuck Berry to Mariah Carey.

Gema, a German royalty-collection group, is suing Suno in a similar case filed in Munich, accusing the service of generating songs that are “confusingly similar” to original versions by artists it represents, including “Forever Young” by Alphaville, “Daddy Cool” by Boney M and Lou Bega’s “Mambo No. 5.”

Major record labels are reportedly negotiating with Suno and Udio for compensation, according to news reports earlier this month.

To detect songs for tagging, Lanternier says Deezer uses the same generators used to create songs to analyze their output.

“We identify patterns because the song creates such a complex signal. There is lots of information in the song,” Lanternier said.

The AI music generators seem to be unable to produce songs without subtle but recognizable patterns, which change constantly.

“So you have to update your tool every day,” Lanternier said. “So we keep generating songs to learn, to teach our algorithm. So we’re fighting AI with AI.”

Fraudsters can earn big money through streaming. Lanternier pointed to a criminal case last year in the U.S., which authorities said was the first ever involving artificially inflated music streaming. Prosecutors charged a man with wire fraud conspiracy, accusing him of generating hundreds of thousands of AI songs and using bots to automatically stream them billions of times, earning at least $10 million.

—Kelvin Chan, AP Business Writer

https://www.fastcompany.com/91355606/how-parisian-music-streaming-service-fighting-ai-fraud?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss

Created 2mo | Jun 20, 2025, 4:10:06 PM


Login to add comment

Other posts in this group

SoftBank’s profit rebounds on AI stocks

Japanese technology conglomerate SoftBank Group Corp. posted a 421.8 billion yen ($2.9 billio

Aug 7, 2025, 4:20:05 PM | Fast company - tech
iOS 26 has political groups panicking over fundraising. The truth might surprise them

Last week, Punchbowl News published an internal memo the National Republican Senatorial Committee (NRSC) sen

Aug 7, 2025, 1:50:07 PM | Fast company - tech
How crypto billionaires took over Trump’s political machine

Last week, President Donald Trump’s super PAC revealed that it has an unsettling amount of cash on hand for a president who is, his

Aug 7, 2025, 11:40:07 AM | Fast company - tech
RushTok is back. TikTok still can’t get enough of sorority recruitment

The internet’s favorite programming is back on: #RushTok season is officially upon us. 

If this is your first time tuning in, “rush” is the informal name for the recruitment process

Aug 7, 2025, 7:10:02 AM | Fast company - tech
Instagram launches map feature. It looks a lot like Snap Map

Location sharing among friends, family, and significant others has quietly become the norm in recent years.

Now Instagram is looking for a piece of the action with the launch of a

Aug 7, 2025, 12:10:05 AM | Fast company - tech
WhatsApp removes 6.8 million accounts linked to scam centers

WhatsApp has taken down 6.8 million accounts that were “linked to criminal scam centers” target

Aug 6, 2025, 9:40:06 PM | Fast company - tech
Google wants you to be a citizen data scientist

For more than a decade, enterprise teams bought into the promise of business intelligence platforms delivering “decision-making at the speed of thought.” But most discovered the opposite: slow-mov

Aug 6, 2025, 7:30:04 PM | Fast company - tech