Announced October 14, 2026, YouTube Music is rolling out labels for tracks where a generative model created the composition and vocals. The tag appears on the player page and expands into the main YouTube app; tapping it explains AI’s role. YouTube says it identifies AI music through music-partner disclosures plus its own detection tools, and that AI involvement alone does not make a track ineligible to earn money.
2026-10-14 · by Moe Ameen
YouTube announced on October 14, 2026 that YouTube Music is rolling out labels for songs made primarily with generative AI, and that the labels will also expand into the main YouTube app. In a company blog post, YouTube framed it as a transparency measure, saying listeners have told it they want to know when they are hearing an AI-generated track.
The label applies to music where a generative model played a significant role — YouTube's example is a track whose composition and vocals were created with generative AI. It shows on the player page; tapping it opens a menu with more detail about what AI did in that specific track. Music where AI played only a minor role may go unlabeled, so the tag is meant to flag primarily-generated songs rather than any use of AI in production.
YouTube says it identifies AI-generated tracks two ways: through disclosures from its music partners and through its own detection tools. On the partner side, recordings are declared under one of three designations — reported by TechCrunch as "Fully Gen AI" (entirely AI-generated), "Partly Gen AI" (AI-generated components alongside the artist's own live elements), and "No Gen AI" (human recordings that only used tools like pitch correction or Auto-Tune). This is a music-delivery classification that flows in through distributors and labels, and is separate from the AI-use question creators answer when uploading a video or Short in YouTube Studio.
Crucially, this is disclosure, not a ban. YouTube has said AI involvement alone doesn't make a track ineligible for monetization, so the label by itself does not demonetize a track. That is a softer stance than some rivals: Deezer labels detected AI-generated music and excludes it from algorithmic recommendations and editorial playlists, and Apple Music has told partners its AI Transparency Tags will surface to listeners as "Made With AI" labels later in the year. Treat the exact on-screen wording and the rollout timing as still settling — YouTube described it as rolling out, and the three-tier category names come from press coverage and distributor documentation rather than a single official on-screen screenshot.
This news is really about one thing creators keep underestimating: platforms are tightening disclosure, and the safe play is to control what goes into your content before a detector decides for you. YouTube Music's labels focus on songs where a generative model made the composition and vocals — the soundtrack layer. That is a useful line to draw when you think about how your short-form video is actually assembled. [Kompozy](/) is a content generation and publishing engine, not a music generator: its video formats — [Persona Shorts](/glossary/persona-shorts) (HeyGen talking-head avatar + native voiceover, optional Pexels b-roll) and Marketing Shorts (a HeyGen hook plus demo footage and music) chief among them — carry narration, not an AI-generated song track that would trip YouTube Music's "primarily AI" label. So the audio bed in a typical Kompozy video is a narration-and-stock composition, which sits on a different side of this specific policy than a Suno-style generated song dropped under a clip.
Where Kompozy earns its place is the step after the policy question: turning one idea into a steady, clearly-yours body of video and publishing it natively to YouTube alongside the eight social platforms plus blog and email, on a schedule, behind a per-post review gate. You stay in control of disclosure because you control the inputs — the persona voice, the [Persona Brief](/glossary/persona-brief) that governs tone, the footage, and whether [Autopilot](/glossary/autopilot) posts for you or a human approves each one. When a platform's disclosure rules shift again (and they will), the content you shipped through Kompozy is a known quantity: generated on brand, persisted as durable [output](/glossary/output-buckets), and labeled by you rather than guessed at by a detector. For the wider picture on AI audio, see [AI-generated music is flooding streaming platforms](/guides/ai-generated-music-flooding-platforms) and, for the practical scoring-and-licensing side, [AI-generated music in promotional videos](/guides/ai-generated-music-in-promotional-videos).
Songs made primarily with generative AI — YouTube's example is a track whose composition and vocals were created by a generative model. The label shows on the player page and, when tapped, explains the role AI played in that track. Music where AI was used only minimally may go unlabeled, so the tag targets primarily-generated songs rather than any use of AI in production.
By YouTube's own statement, not on its own: AI involvement alone doesn't make a track ineligible for monetization. That is different from Deezer, which labels detected AI music and excludes it from algorithmic recommendations and editorial playlists.
Two ways working together: disclosures from its music partners (distributors and labels declare a recording as "Fully Gen AI," "Partly Gen AI," or "No Gen AI," per press coverage) and YouTube's own detection tools. Because detection runs alongside disclosure, a track you did not flag can still be identified and labeled — the open question is detection accuracy and the risk of false hits on human artists.
Not for their soundtrack in the typical case. Kompozy generates short-form video from HeyGen voiceover and licensed Pexels b-roll rather than from an AI-generated song track, so the audio bed is a narration-and-stock composition rather than a "primarily AI" song. You still control disclosure through your inputs and YouTube Studio's own AI-use question when you publish.