AI-generated tracks are no longer a fringe novelty on streaming services. They’re now a structural problem reshaping royalty pools, playlist quality, and listener trust, according to The Verge AI. The Verge AI reports that Deezer is fielding 75,000 AI uploads a day, with machine-made music threatening to overtake human-made uploads on the platform.
The scale shift happened fast. In September 2025, Deezer pegged AI uploads at 28 percent of new music. By year’s end, that climbed past 50,000 tracks daily and 34 percent of uploads. Spotify pulled over 75 million spam tracks in 12 months. What started in 2018 as experimental art projects from Taryn Southern and Holly Herndon turned into an industrial firehose the moment Suno (December 2023) and Udio (April 2024) made full-song generation a one-prompt task.
What stands out here is how quickly the platforms split into two camps on enforcement.
The detection-vs-disclosure split
Deezer went first with actual detection. It labels AI tracks, blocks them from algorithmic recommendations, and has demonetized 85 percent of those streams. Qobuz followed with detection plus an AI charter promising no AI in editorial or curation. CEO Alexis Lanternier framed it bluntly: AI music “is now far from a marginal phenomenon.”
Apple, Spotify, and YouTube took the lighter route: voluntary disclosure. Apple Music “requires” Transparency Tags but defers to providers on what counts as AI. When The Verge AI pressed on enforcement, Apple declined to comment. Spotify is leaning on the DDEX standards body to build industry-wide labels covering lyrics, vocals, and backing music separately, with DistroKid as its first partner. Bandcamp is the only platform with an outright ban on AI-generated music.
Spotify’s Sam Duboff acknowledged the core technical bottleneck: third-party detection tools still produce “a material amount of incorrect assessments.” That’s why the major platforms are hedging on self-reporting rather than committing to automated takedowns.
Listeners aren’t buying it
The demand side is brutal for AI music. A Deezer/Ipsos study found 51 percent of people think AI will produce more low-quality, generic music. The Hollywood Reporter and Frost School of Music polled listeners and got 66 percent saying they never knowingly listen to AI music. More striking: 52 percent said they wouldn’t listen to their favorite artist’s work if they knew AI helped make it.
Singapore researchers point to why. Music is an emotional transaction, and AI output, in their words, may be “perceived as less capable of conveying authentic emotion or fostering meaningful connections.”
Why this matters now
This isn’t just a content moderation story. It’s a royalty-pool story. Streaming services pay out a fixed pot proportionally to streams. Every AI track that captures a play siphons money from human artists. Deezer’s 85 percent demonetization rate is the first real financial countermeasure, and it sets a precedent the rest of the industry will be pressured to match.
The regulatory wedge is also forming. DDEX includes Amazon, Google, Meta, Apple, Pandora, UMG, Sony, and Warner. If a labeling standard locks in there, distributors who don’t comply lose access to the major pipes.
Practical takeaways
- Independent artists: Register with a DDEX-aligned distributor and get the new AI credit metadata right. Mislabeling, even unintentionally, will become a flag.
- Labels and managers: Audit your catalog for AI-assisted production now. Voluntary disclosure today beats forced disclosure later when detection tools mature.
- Platforms and product teams: The Deezer model (detect, label, demonetize, derecommend) is the emerging template. Voluntary-only systems will look thin in 12 months.
- Listeners: Expect “Verified by Spotify” style badges to spread. Trust signals are becoming a product feature.
The next 18 months will decide whether streaming becomes a tagged ecosystem with clear human/AI lanes, or a free-for-all where the royalty math collapses for working artists. Deezer just bet on the first outcome. The rest of the industry is still hedging.
For the full breakdown, head to the original report at The Verge AI.