YouTube Will Now Tell You If a Song Was Made With AI
YouTube Music said on October 14, 2026, it will start labeling songs made with generative AI. The note will appear on those tracks in YouTube Music and in the main YouTube app. TechCrunch
The label sits on the player page, where you press play. Clicking it shows more about how AI was used on that song. The information stays with playback instead of sitting in background files most listeners never open.
YouTube calls the public label "Made with AI". The Verge It covers tracks changed or fully made with AI. Using AI software to create the music and the singing counts for the label.
Behind that label is a three-choice system for music partners: "Fully Gen AI", "Partly Gen AI", and "No Gen AI". "Fully Gen AI" means a track made entirely by AI. "Partly Gen AI" means artists used AI-made parts but added live playing or singing. Songs edited with pitch-fixing tools such as Auto-Tune count as "No Gen AI".
Reporting is required. YouTube requires music partners to report generative AI use. When that report is missing, YouTube says it may use "other signals" to find AI-made songs. That work builds on cooperation with music partners and detection tools already in use to find AI music. YouTube is using the same labeling approach it uses for AI-labeled videos.
AI use alone does not block payment on YouTube. Labeling and payment rules are separate. A track can carry the label and still earn money under existing music rules.
The move follows earlier work on AI tools and rights. On September 23, 2026, YouTube introduced AI creator tools and shopping updates. In August 2023, YouTube started a Music AI incubator with Universal Music to study AI use in music.
The broader context here is keeping track of millions of songs. Like a food label that lists ingredients, a simple tag is easier to sort and check than long notes. Fully, partly, or none fits into databases and upload forms. That reduces confusion when songs are uploaded.
Looking at what this means for trust and tooling, checking becomes the hard part. Reports from partners work when everyone has reason to be honest. Detection tools fill the gaps. No detector is perfect. A live voice over AI-made music will test the line, especially between normal studio help and AI writing. Keeping Auto-Tune outside the rule helps, but questions will remain around fake voices, AI-made parts, and mastering software.
In my view, the hopeful point is simple. Clear labels leave room for mixed work. Artists can use AI for music, arrangement, or singing, add live parts, and still share the song with its history attached. Listeners get context. Platforms get clean data for search and rights. That lasts longer than trying to ban AI audio outright, which has never worked once tools spread widely.
