YouTube Music has begun rolling out a “Made with AI” label for songs that were fully generated or substantially altered with generative tools. The label appears on a track’s player page; tapping it reveals more about AI’s role in the work.¹

The rollout starts in YouTube Music and is due to expand to the main YouTube app. Music partners can designate uploads as “Fully Gen AI,” “Partly Gen AI” or “No Gen AI,” while YouTube says it will also work with partners and use its own detection systems when disclosure is missing.² Its stated aim is clarity: “the right information and clarity” for listeners deciding what they are hearing.²

The policy draws a line between meaningful generation and ordinary production polish. A composition and vocals created by a generative model can trigger the label, but tracks using tools such as pitch correction are classed as non-generative AI.¹ YouTube is also leaning on signals such as SynthID watermarking for Google-made material and potentially C2PA metadata, though detection remains imperfect and an unlabeled song is not proof of human authorship.³

That limitation matters as AI creation tools become more capable. In a repost, Elon Musk highlighted a Grok bot said to use Suno, Midjourney and Opus to make music videos, with the teaser reportedly produced autonomously.Elon Musk (elonmusk)⁴

Other streaming services have chosen tougher guardrails. Deezer labels detected AI tracks, keeps them out of recommendations and editorial playlists, and has pursued a listener filter; TIDAL similarly allows users to disable fully AI-generated recordings. YouTube, by contrast, says AI involvement alone does not bar monetization.¹

Critics of the lighter-touch approach argue the label may be easy to miss during ordinary listening, and YouTube offers no switch to block labeled AI songs from playlists. In that sense, the platform is betting that informed choice—not exclusion—is the workable response to music’s synthetic flood.³