TIDAL is no longer treating fully AI-made tracks as another metadata problem. It is cutting off royalties now, and that puts the money question where it belongs.
Every other large music platform has been trying to label the flood of AI-generated music, contain it, or push it away from recommendation systems. TIDAL has gone further. As The Verge reported on June 29, TIDAL says tracks it identifies as 100% AI-generated will no longer earn royalties, and from July 15 the service will start marking those tracks with a visible AI icon. Music tied to impersonation, deception, high-volume uploads or unusual streaming activity can be removed or blocked outright.
That is the right fight to pick. If you leave the payout system untouched, you leave the machine running. A badge may inform listeners, but it doesn't stop a distributor from spraying thousands of tracks into a streaming catalog and hoping a small fraction of them catch enough plays to skim money from the pool. TIDAL's policy goes after the reason the spam exists in the first place.
The numbers explain why a streamer would stop being polite about this. Deezer said in April that nearly 75,000 fully AI-generated tracks were being delivered to its service each day, about 44% of daily uploads. PC Gamer, citing Deezer's own figures, reported that 85% of streams on those AI tracks were flagged as fraudulent. That isn't a quirky new genre turning up in the catalog. It's industrialized upload behavior, and the cost lands on human musicians because every fake stream competes inside the same royalty economy.
Deezer has already been aggressive by building detection tools, limiting visibility for fully AI-generated music and tagging AI albums. Spotify has moved in a different direction, with The Verge noting that it launched a verification program in April for artists it has confirmed are real people, while profiles that mainly upload AI-generated content don't qualify. Those are useful signals. They are not the same as saying the track won't be paid.
TIDAL's language is careful. The company says royalties should go to original works directly produced, written and performed by people, and it won't knowingly attribute royalties to music it identifies as wholly AI-generated. It also says detection shouldn't be its job alone, so distributors will be expected to label AI-generated music properly. Frankly, that part is overdue. Asking platforms to clean up metadata after the fact while distributors keep feeding the pipe was never going to work.
For Suno and Udio, this is awkward timing. Both companies have spent the past two years trying to move from courtroom defendants to licensed music infrastructure. The RIAA-backed lawsuits filed in 2024 accused Suno and Udio of training on copyrighted recordings without permission. Warner Music Group later settled with Suno and struck a licensing partnership, while The Guardian reported that Warner had also settled with Udio. AP reported in October 2025 that Universal Music Group settled with Udio and announced a controlled AI music platform. Sony's claims against both companies have continued.
Those deals only make full commercial sense if AI-generated music can reach listeners and produce revenue under approved terms. TIDAL's new rule doesn't kill that model across the industry, because TIDAL isn't Spotify and it doesn't control the market. But it does show another possible future: licensed AI tools may be allowed to exist, yet still be denied ordinary streaming royalties when the final track has no human performance, writing or production behind it. That is a much narrower business than the one AI music companies have been selling.
There is a hard technical problem underneath the clean policy language. TIDAL says the July 15 label applies to tracks it identifies as 100% AI-generated, and that it may expand labeling to substantially AI-generated music as tools improve. The company hasn't explained which detection systems it uses, how accurate they are, or what happens when a human artist is wrongly flagged. Those questions aren't paperwork. If you make music with a software-heavy workflow, you need to know where assistance ends and authorship begins.
The industry has never drawn that line neatly. Producers have used drum machines, pitch correction, samples, synthetic instruments and editing software for decades. None of that makes a track fake. A fully synthetic song pumped out at scale by a prompt system is different, but there will be messy cases in the middle, and TIDAL's "living document" framing is an admission that the rulebook isn't finished.
Still, the direction is clear. Governments are moving too slowly to set practical rules for AI music royalties, so platforms are writing the first version themselves. Congress hasn't passed comprehensive legislation for AI-generated content. The EU AI Act deals with transparency, but it doesn't tell a streamer how to split royalties when a track is machine-made. Into that gap comes platform power, and this time TIDAL is using it in favor of human artists.
You should expect the next pressure point to be distributors, not listeners. If TIDAL can make AI labeling and royalty denial stick, Spotify, Apple Music and Amazon Music will have to decide whether they want the same protections or whether they are comfortable paying catalogs that rival services have already marked as synthetic. The money will answer before the philosophy does.
Also read: Ionic Digital takes Celsius Network's ruins to Nasdaq at a $2 billion valuation • A 20-year-old high school dropout just raised $31 million to stop AI data centers from poisoning themselves • Google makes personalized AI image generation free and the pressure on image startups just got real
