Deezer is now receiving 75,000 AI-generated tracks per day, a volume that threatens to overtake human-made uploads on the French streaming service. By the end of 2025, fully synthetic music made up 34% of Deezer's daily uploads, up from 28% in September 2025. Spotify, meanwhile, removed more than 75 million spam tracks over a 12-month stretch, a number that puts the scale of the flood in context.
Generative AI uploads across streaming services have grown nearly 40% in the past year, according to Deezer. Yet listening hasn't kept pace. AI music accounts for roughly 1% of streams on Deezer as of April, up from 0.5% in early November 2025, and 85% of those AI streams are now flagged as fraudulent — up from a previous figure of 70%.
The trigger was the launch of consumer-grade text-to-music tools. Suno arrived in December 2023 and Udio followed in April 2024, collapsing music creation to a prompt box. Before that, AI music was the territory of artists like Taryn Southern, whose I AM AI dropped in 2018, and Holly Herndon, whose Proto followed in 2019, both built on tools like Google Magenta.
Key facts
- 01AI-generated tracks reached 75,000 daily uploads on Deezer, up from 50,000 per day at the end of 2025.
- 02Spotify pulled more than 75 million spam tracks over 12 months and launched AI credits with DDEX, starting with DistroKid.
- 03Deezer has demonetized 85% of AI-generated streams and blocks them from algorithmic recommendations.
- 0466% of listeners say they never knowingly play AI music, per a Hollywood Reporter and Frost School of Music poll.
- 05AI music accounts for roughly 1% of Deezer streams as of April, but 85% of those streams are fraudulent, up from 70%.
Deezer was the first major platform to deploy a detection-and-label system, and it has demonetized 85% of AI-generated streams while keeping them out of algorithmic recommendations. CEO Alexis Lanternier said the company hopes "the whole music ecosystem will join us in taking action to help safeguard artist's rights and promote transparency for fans." Director of Research Manuel Moussallam expects volume to keep climbing: "It is likely that deliveries will keep increasing."
“AI-generated music made up 28% of Deezer uploads in September 2025 and 34% by year-end, with daily volume climbing to 75,000 tracks.”— Jaeden Schafer
Qobuz published an AI charter promising never to use AI for editorial or curation work, declaring that "the heart of Qobuz is and will remain human." Apple Music requires Transparency Tags but relies on self-reporting from labels and creators, and the company declined to detail enforcement when asked, pointing instead to a March industry newsletter saying it "defers to content providers."
Spotify is taking a middle path. It launched AI credits in mid-April with DistroKid as its first partner and is working with standards body DDEX — whose members include Amazon, Google, Meta, Apple, Songtradr, Pandora, BMI, UMG, Sony Music Entertainment, and Warner Music Group — to label whether AI was used in lyrics, vocals, or backing music. Sam Duboff, who leads Marketing and Policy at Spotify for Artists, said third-party detection tools still produce a "material amount of incorrect assessments."
Google requires AI labeling on YouTube and YouTube Music and warns that failure to disclose can lead to removal or suspension from the YouTube Partner program. The company hasn't publicly detailed its detection stack, saying only that it builds on existing spam and clickbait systems. Bandcamp is the lone holdout with an outright ban, though it relies on manual user reports rather than proactive scanning.
Listener sentiment is hostile. A Deezer and Ipsos study found 51% of respondents believe AI will lead to more low-quality, generic music. The Hollywood Reporter and Frost School of Music poll put the share of people who never knowingly listen to AI music at 66%, and 52% said they wouldn't listen to a favorite artist's track if they knew AI helped make it. Researchers in Singapore attributed the bias to emotion: AI music, they wrote, "may be perceived as less capable of conveying authentic emotion."
The counterargument from the platforms is that the line is blurring. Duboff said Spotify expects "the use of AI in music will increasingly be a spectrum, not a binary. Tracks won't be 'categorically AI' or 'not AI at all' with no in between." That framing conveniently lets Spotify and Google avoid the harder question of whether to demonetize AI tracks the way Deezer has, or exclude them from recommendations.
The economics explain the hesitation. Streaming services pay out a fixed pool of royalties divided by share of streams, so every fraudulent AI play diverts money from human artists. Deezer's data — 85% fraudulent streams on AI tracks, concentrated on a handful of viral songs like Velvet Sundown, Breaking Rust, and Solomon Ray — suggests the consumption isn't organic, it's gamed. Yet only Deezer has been willing to absorb the optics of pulling those streams from the payout pool.
The likely outcome is a two-tier system: platforms label AI music, demonetize the worst of the fraud, and let the rest sit in the catalog drawing whatever streams it can. That's not a ban and it's not an embrace — it's a managed decline in the value of a stream, paid for by every artist whose royalty check shrinks as the upload count grows. Until a major platform follows Deezer in cutting AI tracks out of recommendation engines entirely, the flood is the business model.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
