YouTube Music has begun rolling out labels on songs “created primarily using generative AI,” YouTube announced Wednesday. The label sits on the player page, and tapping it opens a short menu explaining the designation. YouTube says the labels will expand to the main YouTube app, and that it’s identifying AI tracks by working with music partners and using its own detection tools.
Most of the machinery runs through metadata. Under YouTube’s Help Center guidance, labels and distributors declare one of three designations when they deliver music through DDEX or CSV templates: Fully Gen AI, Partly Gen AI, or No Gen AI. A track downloaded from a text prompt is fully AI. An artist who generates a bassline or string section and then records live vocals and instruments over it is partly AI. Pitch correction like Auto-Tune and automated AI mastering both count as No Gen AI.
When partners leave the field blank, YouTube says it “will rely solely on those signals and may designate content as fully or partly Gen AI.” The uploaders also do the interpreting, since “it is up to partners to interpret the Gen AI definitions provided by DDEX.”
Two answers are still missing. YouTube hasn’t said whether Partly Gen AI tracks get the visible label, or whether labeled songs are treated differently in recommendations or payouts. Its May update on AI video labels said a disclosure label alone doesn’t change how a video is recommended or whether it can earn money, but that applied to video, not music.
That leaves YouTube as the transparency-only outlier in a market that’s splitting fast, a pattern Zinstrel mapped in AI Labeling Is Everywhere Now. TIDAL stopped paying royalties on fully AI releases in July. Spotify keeps AI Persona artists out of recommendations by default. And one day before YouTube’s announcement, Deezer said listeners will soon be able to strip fully AI music from search entirely. YouTube’s label tells listeners what they’re hearing without, so far, giving them a way to act on it, even though subscribers have complained since January about AI tracks crowding their feeds.
For builders, the definitions matter more than the badge. YouTube has put AI mastering and pitch correction on the human side of the line and generated stems in the middle, which signals which features will carry disclosure weight. It also shifts the burden onto distributors, whose metadata now decides how AI tracks are presented on one of the world’s largest music platforms. An honest, granular AI field at upload is becoming infrastructure.