Every distributor upload form now seems to ask some version of the same question: did you use AI on this? Spotify, Apple Music, YouTube Music and TIDAL each ask it a slightly different way, through different fields, with different consequences. If you make music with AI tools the way we do, one human songwriter working with an AI collaborator, the easiest way to answer honestly and quickly is to keep a simple record as you go. This guide walks through how we'd set one up, how to map it to each platform, and why it's worth keeping even when a field is optional.
One note before we start: platform rules are changing month to month. Everything below was checked in October 2026 against the sources listed at the end. Check your distributor's own help pages before each release, because they decide which fields actually reach each service.
Why Bother Documenting at All?
Three reasons. First, the fields are per part, not per song. Spotify's AI credits apply to individual roles such as lyrics, vocals, instrumental performances and production, not to the whole track. Apple Music's tags split artwork, track, composition and music video. You can't fill those in accurately from memory three weeks after a session. Second, you'll answer the same questions more than once: to your distributor, to a platform form, maybe to a sync client or a collaborator. One record covers them all. Third, detection is now part of the picture. YouTube says it uses its own detection tools alongside what partners declare, and TIDAL says it will tag and stop paying royalties on music it identifies as wholly AI-generated. If a track is ever flagged, a dated record of who did what is the clearest way to explain your process.
Step by Step: Build a Release Evidence File
Step 1 — Make one folder per release. Name it after the track and the release date. Everything below lives in there: a short session log, your lyric drafts, exported files and a disclosure summary. A plain text file or a simple spreadsheet is enough. The point is that it exists and has dates in it.
Step 2 — List every part of the song. Use one row per element so you can map it to any platform later: lyrics, melody / composition, lead vocal, backing vocals, instruments, stems / arrangement edits, mix, master, cover artwork, video / visualiser. These cover every category the four platforms currently ask about.
Step 3 — For each part, write who or what made it. Keep it factual and short. For example: Lyrics — written by [your name], edited after two Suno generations, Lead vocal — AI-generated (Suno), no human vocal recorded, Mix — done by [name] in [DAW], EQ and de-essing on AI vocal stem. If both a person and an AI tool contributed to the same part, say so. Spotify's own guidance is to credit human and AI contributions to the same role separately.
Step 4 — Log the tools and dates as you work. Note the tool name, the model or version shown in the app, and the date of each session that ended up in the final track. Keep the style prompt and lyrics you used for the generation you kept. If you edited stems, note what you changed. This is the part that's impossible to rebuild later, so do it during the session, even if it's one line.
Step 5 — Keep the raw files. Save your original lyric draft (with its file date), the generation you kept, any stems you exported and your final master. For artwork, keep the source image and any edit files. You don't need to show these to anyone unless asked, but they back up what your log says.
Step 6 — Write a one-paragraph disclosure summary. Two or three plain sentences: what was human, what was AI, and which tools. For example: Lyrics and arrangement by [name]. Vocals and instrumental performance generated with Suno, then edited, mixed and mastered by [name]. Cover art: [human / AI-generated]. You'll paste or adapt this into each form.
Map Your Record to Each Platform
Here is how the same evidence file translates into each service's current fields. In almost every case you don't fill these in on the platform itself: the information goes through your distributor or label when you upload.
Spotify — AI credits. AI credits mark specific roles as AI-generated rather than labelling the whole track. Spotify's help page gives lyrics, vocals, instrumental performances and production as common examples. They're optional, currently in beta, and only available through certain distributors. Listeners see them in the song credits and Now Playing view on mobile, not as a badge on the track or your profile. Map your rows for lyrics, vocals, instruments and mix/master to those roles. If your public artist identity is itself AI-generated, Spotify for Artists has a separate AI Persona disclosure.
Apple Music — Transparency Tags. Apple announced four tags in March 2026: Artwork (album level), Track (a material portion of the sound recording), Composition (lyrics or other compositional elements) and Music Video. More than one can apply. Labels and distributors set them, and Apple leaves it to them to decide what qualifies. Apple has said they'll become required for new content in future. Map your artwork row to Artwork, your vocal and instrument rows to Track, and lyrics and melody to Composition.
YouTube Music — AI labels. Since 14 October 2026, partners declare one of three levels: Fully Gen AI, Partly Gen AI or No Gen AI. YouTube's help documentation, as reported, applies this separately to the composition and the vocals. The label shows on the player page, and tapping it explains AI's role. Ordinary Auto-Tune-style pitch correction counts as no generative AI, and using AI doesn't on its own stop a track from earning money. If nothing is declared, YouTube may decide using its own signals, which is another reason to declare it yourself. Your evidence file tells you quickly whether each part is fully, partly or not AI.
TIDAL — AI policy. TIDAL defines AI-generated music as music "wholly or substantially" made by generative AI. Its policy (effective June 2026, last updated 19 October 2026) says it accepts AI music but won't knowingly pay royalties on music it identifies as wholly AI-generated, and shows an icon on content it identifies as 100% AI-generated. It expects distributors to identify AI content before delivery but doesn't name a specific form field. In practice, answer your distributor's AI question accurately and keep your record in case TIDAL's tagging ever disagrees with how the track was made.
Before You Hit Upload
Step 7 — Ask your distributor what it actually sends. A field on an upload form doesn't guarantee it reaches every platform. Ask in writing which AI disclosure fields it delivers to Spotify, Apple Music, YouTube and TIDAL, and save the reply in your release folder. Distributor InterSpace makes the point that you can't type a disclosure into Spotify yourself, so the distributor is the only route.
Step 8 — Decide the level per track, not per artist. One song might have human vocals and an AI backing track. The next might be almost entirely generated. Fill in each release from its own record rather than copying last month's answers.
Step 9 — Keep the folder after release. Platforms are still adding tags and detection, so a record you write today may be useful a year from now. Back it up somewhere other than your laptop.
Common Mistakes
Ticking one box for the whole song: most fields now ask about parts. Over-declaring and under-declaring are both inaccurate.
Forgetting the artwork and video: Apple tags them separately, and a song with human-made audio can still have AI-made cover art.
Leaving it blank and hoping: a blank field doesn't mean "no AI", and YouTube says it may decide for you from other signals.
Writing the log afterwards: a log rebuilt from memory is weaker than one-line notes made during each session.
Treating today's rules as fixed: rules and categories are still moving. Re-check before each release.
Where That Leaves Us
Disclosure takes a few minutes per release if the notes already exist, and a frustrating hour if they don't. We think being open about the AI side of how songs are made is simply the right way to do this, and a tidy evidence file makes it painless. If you'd like to hear how human writing and an AI collaborator sound together, our songs are on the homepage, and our Suno Guide collects the prompts, style tags and how-to guides we use.
Sources
Spotify for Artists — AI credits (per-role credits, beta, via distributors, where they show)
Music Business Worldwide — Apple Music launches AI transparency tags (the four tags, who declares, March 2026)
TechCrunch — YouTube Music launches new labels for AI-generated music (three levels, player-page label, monetisation)
Tech Insider — YouTube Music AI song labels (composition and vocals declared separately, undeclared tracks)
TIDAL — AI policy (definition, royalties, tagging, distributor expectations)
InterSpace — Spotify AI credits only travel through your distributor (ask your distributor what it delivers)
