Deezer says over half its uploads are AI-made

Deezer says more than half the songs uploaded to its platform are now AI-generated, and it’s responding by removing fully AI-made tracks used for fraudulent streaming schemes. The French music service said Tuesday that it will also delete tracks that have gone six months or more without a stream.

The company says it is receiving nearly 90,000 fully AI-generated tracks every day, after identifying more than 13.4 million AI tracks on its platform last year. In June 2025, Deezer started tagging albums that include AI and excluding AI music from its algorithmic and editorial recommendations.

While fully AI-generated songs account for just 1% to 3% of total streams on Deezer, the company said 85% of those streams were fraudulent last year. Deezer also said it does not pay royalties on streams it determines were manipulated.

How fake AI music streaming works

Deezer’s move targets a broader scam affecting services including Spotify, Apple Music, Amazon Music and Tidal. Using tools such as Suno, Udio and AIVA, fraudsters can generate thousands of tracks quickly, upload them in bulk, then use bots and streaming farms to inflate play counts and collect royalty payments.

Earlier this year, a North Carolina man pleaded guilty to a scheme like this after collecting more than $8 million in royalties. Billboard reported in November that Suno’s 2 million-plus users are generating 7 million AI songs every 2 weeks — more than the entire Spotify catalog. A recent study estimated creators could lose nearly 25% of their revenue, or more than $4 billion, by 2028 because of fraudulent AI music streams.

Recommended reading

Entertainment Apps Are Converging Around AI

Apple Music VP Oliver Schusser told The Hollywood Reporter in February that his division detected 2 billion fraudulent streams in 2025, blocking $17 million in royalty payouts.

Spotify and Tidal are adding labels too

Other platforms have started adding their own safeguards. In late April, Spotify said it would add verification badges to artist profiles to help distinguish human-made music from AI output. Last month, Tidal said it would label fully AI-generated music and remove fraudulent or deceptive tracks.

CNET AI reporter Katelyn Chedraoui said labels and badges alone won’t solve the problem.

“Streaming giants will have to be extremely vigilant, but so far, they haven’t cracked down hard on AI music. So, the responsibility to avoid AI music slopification lies with us. People who are uploading AI-generated or AI-assisted music ought to be transparent with listeners.”

CNET senior editor Kourtnee Jackson said that until regulation arrives, listeners will largely be left to figure it out themselves.

“Consumers can generally just rely on labels, badges and word-of-mouth to discern what’s made with AI and what’s not. Be aware that independent creators, musicians and companies like Suno are contributing to the creation and consumption of AI-generated music.”

In a press release, Deezer CEO Alexis Lanternier said the company wants to protect artists and songwriters while keeping the service focused on “music that fans actually love.” He also urged the industry to align on common rules.

“If the industry can align and come together to create shared standards, we can build a healthy and sustainable music ecosystem, where great music wins.”