The next song you queue up on YouTube Music might come with a small honesty sticker attached. From Wednesday, 14 October 2026, the platform has begun rolling out labels that tell listeners when a track was created primarily with generative artificial intelligence — placing AI disclosure right on the player page, in the one spot almost every listener actually looks at.
Tapping the label opens a short explanation of what the AI actually did: whether it composed the music, generated the vocals, or both. It is a transparency feature, not a ban. AI-made tracks will still play and still earn money — but listeners will be able to see, before deciding how to engage with a song, that a machine played a significant creative role in making it.
"Generative AI is opening up new ways for artists to create, experiment, and share music," YouTube said. "As these tools evolve, our listeners have told us they want transparency when they're hearing an AI-generated track." The labels will also expand to the main YouTube app, making AI disclosure a shared feature across both platforms.
How the labels work
When a track qualifies, a badge appears directly on the player page, in the Now Playing view near the song title and artist name. It is described as a compact "AI" marker with an info icon; tapping it opens a "How this was made" panel explaining the AI's role, with a link to more detail.
The trigger is not any use of AI at all. The label applies when generative models handled core creative work — composing the music, producing the vocals — rather than acting as a minor production aid. A song that merely passed through AI-assisted mastering or editing would not normally cross the line.
YouTube identifies candidate tracks two ways: disclosures from music partners and its own detection systems. Partners declare AI use when delivering tracks through DDEX or CSV templates, and YouTube's automated tooling cross-checks what arrives. The company says it is applying the same labeling principles already used across the main YouTube platform.
Fully, partly or not at all: the three designations
According to YouTube's help documentation for music partners, every delivered track can now carry one of three designations:
- Fully Gen AI — the entire track was generated by AI: a text prompt typed into an AI music generator, downloaded as a finished track, or a composition, instrumental and vocals all created with a generative model.
- Partly Gen AI — some of the content was created with generative AI. This is the broadest category, covering AI-generated basslines or string sections, lyrical themes, co-written verses, and melodies later recorded in a studio.
- No Gen AI — none of the content was created with generative AI.
Two details in the "Partly" category deserve attention. First, the designation covers the content connected to a track, not just its audio: a song performed entirely by humans can still be labelled if its accompanying visuals were produced by a text-to-video AI model. Second, YouTube has drawn a firm line around conventional production tools — tracks edited with pitch-correction software such as Auto-Tune are explicitly classed as "No Gen AI".
That Auto-Tune carve-out matters, because it shows what the label is really for: disclosing when AI made meaningful creative decisions — writing the melody, choosing the words, singing the vocal — not inventorying studio software. Producers who generate a chord progression with AI and then build a human record on top will test that boundary constantly.
Why now: the flood of synthetic songs
Generative music tools have collapsed the cost of producing a finished-sounding track to nearly nothing, and the platforms' royalty pools are paying the price of abundance.
Deezer has become the industry's unofficial canary in the coal mine. The French service reported in July that it was receiving nearly 90,000 fully AI-generated tracks every single day, and that more than half of all uploads to its platform were AI-generated. Much of that output exists not to be loved by listeners but to game recommendation systems and siphon royalty payouts — a phenomenon that shades into streaming fraud.
That context explains the industry's split personality. A machine-made lo-fi beats playlist harms nobody directly — except the human musicians competing for the same stream counts. But when fake streams inflate royalty claims, the money going to synthetic tracks comes out of the pool shared with real artists. Listener sentiment appears to be souring too, which is why every major platform is now racing to be seen doing something.
How it compares to Spotify, Apple and Deezer
YouTube's approach — disclose, but do not penalise — places it in the middle of an emerging industry spectrum.
- Spotify already displays AI labels on some tracks and runs a music spam filter to combat fraudulent uploads.
- Apple Music plans to introduce "Made With AI" labels as part of an "AI Transparency Tags" feature later this year.
- Tidal labels fully AI recordings and gives listeners an option to hide them entirely.
- Deezer has taken the strictest stance: it labels detected AI music, excludes it from algorithmic recommendations and editorial playlists, withholds royalties from fraudulent streams, and is working on a filter to let users block AI music entirely.
The philosophical divide is clear. Deezer treats AI music as a fraud and quality problem to be suppressed. YouTube treats it as an information problem to be disclosed. YouTube's position avoids the near-impossible task of deciding which AI music is legitimate art and which is spam; its risk is that disclosure without consequences changes nothing. Fraudsters do not care about labels, and listeners who do not care will keep streaming.
Where YouTube may have an edge is distribution. Because the same labeling principles already run on the main YouTube app, extending them to music video uploads there closes a gap audio-only streamers cannot match — the millions of AI-generated songs that live on YouTube as videos with static images, not as official music tracks.
YouTube's long road to AI disclosure
This is not YouTube's first attempt at taming synthetic media. In 2023, it introduced disclosure rules for AI-generated video content and launched a Music AI Incubator with Universal Music Group — an attempt to design generative music tools with rightsholders rather than against them. Viral AI fakes imitating major artists, such as the 2023 "Heart on My Sleeve" track with synthetic Drake and The Weeknd vocals, then forced the industry to treat deepfake music as a rights and identity issue, prompting YouTube to offer artists a request-and-removal process for voice clones.
The new song labels are the natural next step in that arc: from reactive takedowns of the worst abuses to proactive disclosure for everything else. Instead of asking whether an AI track should exist, YouTube is now asking whether the listener knows what they are hearing.
