YouTube Music begins labeling songs made with generative AI
publishers 26articles 27first reported 14 Oct, 17:01 UTCdeveloping since 14 Oct · 8 editions
YouTube Music is rolling out labels for tracks created entirely or significantly altered using generative AI, giving listeners more information about how AI contributed to the music.
The label appears on a track’s player page, and tapping it opens a menu with additional details. It covers fully AI-generated songs as well as tracks in which generative models played a major role, such as creating the composition or vocals. Routine pitch correction and similar editing tools are categorized as not generative AI, according to YouTube’s guidance for music partners.
YouTube said it is identifying affected tracks through disclosures from music partners and its own detection tools. Partners can classify uploads as fully generated by AI, partly generated by AI or not generated by AI. If an uploader does not disclose AI use, YouTube may rely on other signals when deciding whether to apply a label.
The system is available first in the YouTube Music app and on its website, with labels expected to expand to music content in the main YouTube app later. YouTube is emphasizing disclosure rather than banning AI-generated music; the use of AI alone does not make a track ineligible for monetization.
Other streaming platforms have introduced their own policies for AI music, ranging from labels to filters and restrictions on recommendations or royalty attribution.
HOW THIS STORY WAS MADE
Written from 7 articles, 4 with the publisher's own text; 7 independent newsrooms once syndicated copies count as one; this version written 24 h after the record first saw the story; the editor kept it as written.
- Publishers
- 26
- Source articles
- 27
- Given to the writer
- 7, 4 with the article's own text
- Independent newsrooms
- 7
- This version written
- 24 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and published it as written.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →