YouTube Music will not attach an AI label to a song merely because an engineer touched up a vocal with Auto-Tune or a producer ran the final master through automated AI mastering software. Both scenarios fall within the "No Gen AI" category the platform established for its music partners, as part of the AI disclosure labels YouTube Music began rolling out on October 14, 2026. The distinction separates AI used for production touch-ups from AI that generates core elements of a track — the line the disclosure system is built around.

The classification examples appear on YouTube's help page. Standard pitch-correction software such as Auto-Tune, used "to correct minor tuning mistakes in a live vocal performance," lands in the No Gen AI class. The same applies when automated AI mastering software handles a track's final balance, compression, and equalization.

A fully human song can still be flagged as partly generated

The Partly Gen AI category covers an artist who uses an AI tool to create a bassline or string section, then layers live singing and instruments on top. It also extends to artists who turn to AI for lyric ideas, co-write verses with others, or sketch out melodies before entering the studio.

Notably, a song with 100% human-performed audio still qualifies as Partly Gen AI if a text-to-video model produced the music video's visuals.

Fully Gen AI applies when someone types a prompt into an AI music generator and downloads the track it returns. The same tag covers an artist who has a model generate the composition, instrumental performance, and vocals — even after multiple prompts to refine the output.

The label targets songs "where AI played a significant role," according to YouTube's announcement. When users tap it on the player page, a menu opens explaining how AI was used, putting the disclosure in front of listeners at the point of playback.

Partners declare AI use through DDEX or a CSV column

If a partner declares no value, YouTube treats the track as unknown "unless we are able to make a determination based on other signals." The company said it works with music partners and has its own tools to identify tracks made by AI, applying the same labeling rules already in use across YouTube.

Partners delivering content through DDEX ERN 4.3.2 or later, or through CSV delivery templates under their SRAV terms, declare whether AI was used. Artists who upload through YouTube Studio will mark it in the attribute area.

The CSV route includes a "contains_ai" column that accepts four values: all, partly, none, or unknown. The "unknown" value clears a value submitted in error. YouTube leaves it to partners to decide how DDEX definitions apply to their own catalogs.

The labels are also coming to the main YouTube app, extending the disclosures to listeners beyond the dedicated music service. AI involvement alone does not make a track ineligible for monetization, TechCrunch reported.

A broader industry shift toward AI transparency

YouTube's Dream Track tool already allows select creators to craft soundtracks using AI versions of participating artists' voices. Deezer, Spotify, Apple Music, and TIDAL all run or are planning their own AI labels, with Deezer saying in July that it was close to 90,000 fully AI-generated tracks a day — a figure that illustrates the volume of fully AI-generated music arriving on streaming platforms.

In September 2026, Cryptopolitan reported that Suno rebuilt its model lineup on Warner Music Group and BMG catalogues. Last year, Spotify collaborated with the major labels to develop AI products after removing more than 75 million spam tracks.