YouTube Will Label AI-Generated Songs, Starting With YouTube Music

By Daniel Okafor

YouTube began labeling AI-generated songs on YouTube Music on Wednesday, and the markers will soon expand to the main YouTube platform. The labels will apply to tracks the company deems were "created primarily using generative AI," based on its own detection tools and those built by its music partners.

"Generative AI is opening up new ways for artists to create, experiment, and share music," YouTube said in a blog post on Wednesday. "As these tools evolve, our listeners have told us they want transparency when they're hearing an AI-generated track."

That second sentence is doing a lot of work. It frames the policy as listener service rather than rights management, which is a polite way of navigating an argument the music industry has already made for it.

Who else is labeling AI music?

YouTube arrives late to a crowded announcement schedule:

• Apple Music said in August it would add "Made by AI" labels to songs later this year.
• Spotify, just before that, said it would label the profiles of apparently AI-generated artists as "AI Personas."
• Tidal has promised both to label AI music and to bar those tracks from accruing royalties.

Tidal's royalty ban is the outlier here. Everyone else is naming the machine; Tidal is also declining to pay it.

What did the industry groups demand?

The industry did not wait for the platforms to act. Over the summer, a coalition representing musicians and the wider music business — including the RIAA, IFPI, the Grammys and SAG-AFTRA — banded together to advocate for a new labeling program indicating whether a song was produced with AI, urging the use of "AI-Generated" or "AI-Assisted" indicators.

"Fans want to know whether and how generative AI has been used in the music to which they listen," International Federation of the Phonographic Industry (IFPI) CEO Vikki Oakley and Recording Industry Association of America (RIAA) chairman and CEO Mitch Glazier said in a joint statement.

What does the labeling actually change?

The gap between "AI-Generated" and "AI-Assisted" is where the policy gets slippery, since almost all commercial music now passes through some machine assistance — pitch correction, mastering algorithms, recommendation-driven production choices. YouTube's trigger, "created primarily using generative AI," pushes the question of degree back onto its detection tools and those of its partners, none of which have been publicly described.

The rollout across the main YouTube platform is next. Whether the labels inform listeners or simply launder the controversy into a metadata field will depend on detection accuracy the company hasn't yet detailed.