Key Highlights:
- Deezer now receives 60,000+ fully AI-generated tracks daily, representing 39% of all uploads
- Up to 85% of streams on AI-generated tracks are fraudulent, up from 70% in September 2025
- The platform detected 13.4 million AI-generated tracks throughout 2025
- Deezer is licensing its detection technology to rival platforms and music companies
- 79% of UK songwriters now worry about AI music competing with human work
The Numbers Behind the Crisis
Deezer’s latest data reveals a streaming fraud epidemic that threatens artist royalties across the industry. AI-generated tracks now account for 1-3% of all streams on the platform. That’s triple the 0.5% share recorded in September 2025.
The fraud rate tells the real story. Bots stealing from artists has become an industrial-scale operation. Deezer confirms that up to 85% of these AI-music streams come from automated bots designed to siphon royalties from the platform’s payment pool.
The upload volume has exploded. At the start of 2025, roughly 10,000 AI tracks hit the platform daily. By November, that figure reached 50K daily uploads on Deezer. Now it stands at 60,000 tracks per day.
How the Fraud Works
Modern streaming fraud has evolved beyond simple bot plays on single tracks. Fraudsters now use AI tools like Suno and Udio to mass-produce thousands of unique songs. Each track receives only a few thousand streams, staying below detection thresholds.
The takeaway: This “dilution fraud” spreads billions of fake streams across massive catalogs, making individual manipulation harder to spot.
The 2024 federal case against Michael Smith exposed the scale of these operations. Smith allegedly generated over $10 million in royalties by uploading hundreds of thousands of AI songs and using 1,000+ bot accounts to stream them. His strategy relied on volume. A billion fake streams spread across tens of thousands of tracks draws less scrutiny than the same number concentrated on one song.
Streaming farms and AI fraud networks exploit the pro-rata royalty model used by most platforms. Every fraudulent stream directly reduces payments to legitimate artists. The math is simple: fake plays dilute the royalty pool that real musicians depend on.
Deezer’s Response
Deezer has built detection systems that identify both AI-generated content and fraudulent streaming patterns. The platform doesn’t recommend or playlist AI tracks. It demonetizes streams flagged as bot activity.
“We’ve seen a great interest in both our approach and our tool, and we have already performed successful tests with industry leaders, including Sacem,” said CEO Alexis Lanternier. “From now on, we are licensing the tech to make it widely available.”
The Deezer detection tool licensing announcement marks a shift toward industry-wide cooperation. Other platforms and music companies now have access to the same AI tagging and detection technology Deezer developed internally.
Deezer’s “”artist-centric”” payment model adds another layer of protection. Artists with 1,000+ monthly streams and 500+ unique listeners receive double royalty weighting. Functional audio like white noise gets demonetized entirely. These changes aim to redirect money toward professional creators and away from AI music catalog clutter.
The Human Cost
Credit: AI-generated with Ideogram
Musicians are watching these developments with growing alarm. A PRS for Music AI survey of 2,600+ UK songwriters found that 79% worry about AI-generated music competing with human work. That’s up from 74% in 2023.
The percentage who believe AI will negatively affect their livelihoods rose from 69% to 76% over the same period. These shifts coincide with increased understanding of the technology. In 2023, 53% of creators said they understood how AI worked in music creation. Now 72% do.
Musician concerns about AI extend beyond streaming fraud. A 2026 survey of 1,100 producers found 77% worried about losing originality in the market. And 93% view AI training on copyrighted works without permission as a violation of their rights.
What This Means for You:
The 85% fraud figure signals a turning point. Streaming data is becoming unreliable for measuring real fan engagement. Talent buyers and promoters who book artists based on inflated numbers risk empty venues.
Deezer licensing AI detection to competitors suggests the industry recognizes this threat requires collective action. The C2PA standard for “Content Credentials” offers another path forward. Major audio stakeholders including the RIAA, Roland, and Avid have joined the coalition to create verifiable proof of human authorship.
For working musicians, the message is clear. Build direct relationships with fans through email and community platforms. Document your creative process to prove human origin. Avoid any service promising guaranteed streams.
The streaming economy is splitting in two. AI-generated “service music” will face increasing demonetization. Human-created work with verifiable authenticity will command premium value. Your strategy should reflect which side of that divide you want to occupy.”
