RTS.FM
← blog

22 JUL 2026 · rts.fm editorial

Deezer Says Over Half of Daily Uploads Are Now AI, Inside Streaming's Flood Problem

Deezer says AI tracks now top 50% of daily uploads, mostly bot-driven royalty fraud, and it's now selling its detection tech to rival platforms.

Deezer says AI-generated tracks now account for more than half of everything uploaded to the platform each day. In April 2026 the company measured AI uploads at 44% of the daily total. By July 2026 that figure had crossed 50%, roughly 90,000 AI-generated tracks arriving on the platform every single day, according to Deezer's own newsroom and reporting from TechCrunch. Deezer says the great majority of streams on fully AI-generated tracks are fraudulent, driven by bots rather than real listeners, and the platform now demonetizes flagged AI content before removing it outright.

This isn't a hypothetical problem for anyone releasing music independently. It's the platform math that decides how a listener actually finds a new release, and right now that math is being flooded.

How bad is the AI upload problem, really?

Deezer's numbers give a rare, platform-reported look at the actual scale of AI content flooding a major streaming service. The climb from 44% in April to over 50% by July 2026 is not a one-off spike, it's a trend line moving in one direction over just a few months. At roughly 90,000 AI tracks uploaded per day, Deezer is describing a volume of machine-generated content that now outpaces everything made by human artists combined on its own platform, at least by raw upload count.

The fraud angle is what makes this more than a content-moderation headache. Deezer has previously found that up to 85% of streams on fully AI-generated tracks were fraudulent, meaning bots, not people, were generating the plays. That matters because streaming royalty pools are typically calculated on a pro-rata basis: total revenue gets divided based on each track's share of total streams across the platform. Every fraudulent stream on an AI track dilutes the pool that real artists, including independent techno and house producers putting out actual releases, are paid from. It's not just spam cluttering search results, it's a direct drain on royalty payouts across the entire platform.

What is Deezer actually doing about it?

Deezer built an in-house AI detection tool that it says operates with 99.8% accuracy, and its enforcement policy is now a two-stage process: flagged AI-generated tracks are first excluded from Deezer's algorithmic and editorial recommendations and demonetized, then removed from the catalog entirely once confirmed. The company has framed this as protecting royalty payouts for the artists actually making the platform's real content.

What's notable is that Deezer isn't keeping this technology to itself. As TechCrunch reported back in January 2026, the company began commercially licensing its detection tool to rival streaming platforms earlier in the year, well before the July numbers came out. That's an unusual move for a company competing directly with Spotify, Apple Music and others for subscribers, but it signals Deezer sees the AI flood as an industry-wide threat to trust in streaming platforms generally, not just a problem it can solve for its own catalog while competitors ignore it. If AI-generated fraud is depressing royalty payouts across the industry, and listeners start to distrust what they're hearing on any platform, that's a shared problem worth selling a shared fix for.

How does this compare to what Spotify and Apple Music are doing?

Deezer's approach, detect, demonetize, then delete, sits at one end of a spectrum of platform responses to AI music. Spotify has taken narrower, more targeted steps: removing tracks that impersonate real artists' voices without consent, and cracking down on AI-generated tracks used specifically for streaming fraud schemes (bot networks that upload thousands of short AI tracks purely to farm royalties through fake plays), rather than banning AI-assisted music outright. Apple Music has been comparatively quiet in public about a formal AI-detection policy of its own.

Bandcamp, by contrast, took the most aggressive stance of any major platform when it banned AI-generated music outright in January 2026, refusing to allow tracks "generated wholly or in substantial part by AI" onto the platform at all, enforced through human review rather than automated detection. That puts Deezer's detect-and-demonetize model in the middle: more proactive than Spotify's fraud-focused approach, but less absolute than Bandcamp's outright ban.

What does this mean for independent electronic labels?

For a small label putting out techno, house or minimal releases, the practical concern isn't abstract. It's discoverability and trust. If a platform's new-release and recommendation algorithms are wading through tens of thousands of AI-generated tracks a day, human-made records risk getting buried under sheer volume, at least until detection and removal keep pace with upload growth. Deezer's demonetization-first approach at least means flagged AI tracks stop earning from streams quickly, rather than sitting in the royalty pool diluting payouts for months while a slower review process runs.

There's also a trust dimension that matters for a label's relationship with listeners. Part of what independent, artist-run labels sell isn't just the track, it's the certainty that a real person made it, in a real studio, as part of a real creative practice. Platforms that can credibly claim to police AI-generated fraud protect that distinction. Platforms that can't risk becoming places where listeners can no longer tell the difference, which is bad for everyone selling human-made music, not just the platform's bottom line.

Deezer's willingness to license its detection tool to competitors suggests the company thinks this fight only works if it's fought industry-wide. Whether Spotify, Apple Music and others adopt something similarly aggressive, or keep responding narrowly to fraud and impersonation cases as they surface, will likely shape how much AI-generated content independent artists are competing against on each platform going forward.

quick wins

  • AI-generated tracks made up 44% of Deezer's daily uploads in April 2026 and crossed 50% (about 90,000 tracks a day) by July 2026.
  • Deezer has found up to 85% of streams on fully AI-generated tracks were fraudulent, bot traffic that dilutes the royalty pool paid out to real artists.
  • Deezer's enforcement is detect, demonetize, then delete, using an in-house tool it claims is 99.8% accurate, which it has licensed to rival platforms since January 2026.
  • Bandcamp banned AI-generated music outright in January 2026; Spotify targets impersonation and fraud specifically rather than banning AI-assisted music broadly; Apple Music has stayed largely quiet on a formal policy.
  • For independent labels, the practical risk is discoverability: AI upload volume can bury human-made new releases in recommendation feeds until platform-wide detection catches up.

The scale Deezer is now reporting publicly makes this less a future risk for independent music and more a live condition of the platforms every small label already depends on. How fast the rest of the industry follows Deezer's lead will decide how much of a problem it stays.

Loading sets…