What Apple Music actually announced
Apple Music will tag songs identified as AI-generated with a small marker on the track listing. The system is designed to flag tracks where the core musical content was produced by a generative model rather than by a human artist or producer. Tracks that use AI as a tool inside an otherwise human-led workflow are not the target. The line, as with Bandcamp's policy a month earlier, is on authorship rather than tooling.
How the detection works
Apple has not published the exact methodology, but the system combines acoustic fingerprinting with metadata signals: distributor, account history, release velocity, and patterns common to bulk AI-generated catalog uploads. Tracks identified at high confidence are tagged automatically. Tracks at the boundary go through human review. Artists can appeal a tag through their distributor.
What gets tagged, and what does not
- Fully generated tracks from text prompts: tagged.
- AI vocal clones of unauthorised artists: tagged and likely removed.
- Mass uploads from accounts producing dozens of tracks per week: tagged.
- Stem separation, pitch correction, mastering assistants: not tagged.
- AI-assisted lyric drafting, with human writing on top: not tagged.
- Producers using AI inside a DAW: not tagged.
The line is on authorship rather than tooling.
What this means for artists who use AI tools
If your workflow involves AI as a production accelerator, this changes nothing. The tag is reserved for songs where the music itself is generated. The risk most working artists run is not their workflow. It is their distribution chain. If you release through a platform that also distributes mass AI catalog, your music can get caught up in batch reviews aimed at the bulk uploaders. The fix is to be deliberate about who distributes your music and how clean your account history looks.
The bigger picture
Apple Music's tag does not ban AI music. It informs the listener. That is a milder position than Bandcamp's outright ban, but it might end up being more consequential. A tagged track is downgraded in recommendations, in editorial eligibility, and in the trust signal that listeners use to decide whether to add a song to their library. The tag is not a punishment. It is a market sorting mechanism, and it will quietly redistribute streams toward artists whose music is verifiably authored.
What to do about it
Three actions that protect the work: keep your sessions, your stems and your authorship trail in a place a distributor can verify on demand. Use a distributor that has a clear position on AI content. Make the human contribution to your music visible in the song notes, the credits and the press materials. The artists who can tell their own authorship story without ambiguity are the ones who benefit when the platform sorts the catalog.
