On the same day this week, the two biggest music-streaming responses to AI-generated songs landed — and they could not be more different. YouTube Music started labeling AI-generated songs on its player page. Deezer announced it will let listeners filter them out entirely before the end of the year. One platform chose transparency. The other chose an off switch.
That makes this the perfect moment to do what nobody else has: rank the whole industry’s answers on one spectrum, from strictest to loosest, and see what each choice reveals.
YouTube Music: Labels, But No Off Switch
YouTube Music’s system has three tiers on the player page: “Fully Gen AI,” “Partly Gen AI,” and “No Gen AI.” Tapping a label opens a “How this was made” panel, and the labels are rolling out to the main YouTube app as well. Partners declare AI use through DDEX or CSV delivery templates, and YouTube layers its own detection tools on top.
The definitions are broader than you might expect. “Fully Gen AI” means a prompt-into-generator track, or a genAI-made composition with genAI vocals. “Partly Gen AI” covers AI basslines and strings, lyrical brainstorming, co-writing, and even AI visuals on a human-performed song. And “No Gen AI” is the interesting one — it still allows pitch correction like Auto-Tune, which is, of course, technology too.
Here is the catch: there is no filter, no hide option, no down-ranking. AI involvement alone does not block monetization either — labeled AI songs keep earning royalties like everything else.
Deezer: the Off Switch
Deezer’s toggle, arriving before year-end, goes much further. Turn it on and AI-generated tracks, albums, and related artists vanish from search results. CEO Alexis Lanternier said the company has seen “clear user demand for the ability to filter out AI-generated music” — a 2025 Deezer survey found roughly 45% of customers wanted AI music removable.
The scale explains the urgency. AI tracks went from 10% of Deezer’s daily uploads in early 2025 to more than 50% — roughly 90,000 per day — by July 2026. The platform already labels detected AI tracks, excludes them from algorithmic recommendations and editorial playlists, withholds royalties from fraudulent streams, and even ships a free online AI-music detector that works across 20 platforms in 27 languages.
The Strictness Spectrum
Ranked from strictest to loosest, the industry now looks like this:
1. Bandcamp — banned AI tracks outright. The nuclear option.
2. Tidal — pays nothing for fully AI-generated music. Artists can upload; they just cannot earn.
3. Deezer — labels plus the listener-controlled off switch, algorithmic exclusion, and royalty withholding for fraud.
4. Spotify — tags AI tracks and artists (“AI Personas”) and runs a music spam filter, but stops short of user filtering.
5. YouTube Music — disclosure labels only. No hiding, no down-ranking, no monetization impact.
6. Apple Music — “Made With AI” tags planned for later this year, still the least-defined response.
Why this matters
The spectrum is really a fight about who the music business serves. Bandcamp and Tidal chose the artist. Deezer chose the listener — betting that 90,000 AI tracks a day is a discovery problem its users will pay to solve. YouTube Music chose the catalog, keeping everything labeled and everything monetized, because its business model runs on volume.
Watch which end wins. If Deezer’s toggle becomes a marketing advantage — the streamer that lets you opt out of the flood — the labels-only platforms will face real pressure to follow. The AI-music era is no longer about whether the songs exist. It is about who gets to decide whether you hear them.
Meanwhile in Apps & Tools: YouTube’s custom feeds — the same company handing you more algorithmic control elsewhere.
Meanwhile in Apps & Tools: Google Play’s private vibe-coded apps — another week, another AI flood Google has to manage.
