Free updates for the book
Making Music with AI — Updates
AI music tools, platform rules and the law change every month; a printed page cannot. This page lists the changes by date, newest first, each naming the chapter it affects. It is reviewed monthly, and sooner when a major decision lands.
Dated changes
· All chapters
Updates page opened. The book was last verified in October 2026. No changes to report yet: every tool, rule and law in the book stands as printed.
What will appear here: law and policy (court decisions, settlements, reports, new laws; rule changes at Spotify, Apple Music, YouTube, Deezer, Bandcamp and the major distributors), tools (new models, renamed products, plan, download and licence changes) and corrections. Every entry is dated and gives its source. Check the updates for Chapters 11 and 12 before any release.
Like the book, this page is general information, not legal advice.
Printable forms
The forms from the book, ready to print. Each one prints on its own page. Use the button on a form to print just that one, or your browser's print command to print them all.
- Idea card
- Prompt template
- Iteration log
- Voice consent checklist and record
- Release checklist
- Record sheet
- Disclosure note
- Personal AI policy
Idea card
From Chapter 3: Song Generators, Used Well
Five minutes, before you open any generator. End with a question the generator can help you answer.
See the book's example
Title: Kitchen Light
About: the hour after everyone has gone to bed, when the house is finally quiet and you can think
Feeling: tired, tender, a bit relieved
Have: a chorus line, "leave the kitchen light on," and the chords D, G, D, A
Want to hear: does this work as a slow waltz, or is it better straight?
Reusable prompt template
From Chapter 4: Prompt Craft for Music
Copy it for each new song and fill in the blanks. Leave a blank empty only on purpose.
[GENRE + ERA], [TEMPO] bpm, [FEEL: straight / swung / half-time], [KEY].
[INSTRUMENT 1 + how played], [INSTRUMENT 2 + how played], [INSTRUMENT 3 + how played].
[SPACE: dry / roomy / wide / lo-fi].
[VOCAL: range, delivery, placement] or Instrumental.
[ARRANGEMENT: what changes over time].
Exclude: [DEFAULTS YOU DO NOT WANT].
Iteration log
From Chapter 4: Prompt Craft for Music
Write what you changed and what happened, one row per batch. Change one layer at a time.
| Run | What I changed | Runs | What I heard | Keep? |
|---|---|---|---|---|
See the book's filled-in example
| Run | What I changed | Runs | What I heard | Keep? |
|---|---|---|---|---|
| 1 | Base prompt: indie folk, 92 bpm, D major | 4 | Tempo near 92 in 3 of 4. Strings in all 4. | No |
| 2 | Added "dry, no strings" at the end | 4 | Strings gone in 2 of 4. Still roomy. | Maybe |
| 3 | Moved "dry, close-miked" to the front | 4 | Close and intimate in 4 of 4. Kept the brushes. | Yes, run 3c |
Voice consent checklist and record
From Chapter 7: Voices, Cloning and Consent
Before any voice work that involves anyone other than yourself. If you cannot tick every box, stop.
You can
- Clone and use your own voice, from recordings where you are the only voice.
- Use another person's voice when you hold a written license covering this exact use.
- Use singing-synthesis voices and royalty-free library voices within their licenses.
- Use licensed artist voices through a platform that handles the artist's approval and payment.
- Describe a vocal style in a prompt without naming a person.
You cannot
- Clone or convert to a real person's voice without their written permission.
- Rely on "they won't mind," a verbal yes, or a public figure's general fame.
- Release music that sounds like a named artist and lets listeners believe it is them.
- Use a deceased artist's voice without the estate's consent.
- Keep using a voice after the license has ended.
Voice consent record
Voice belongs to: Date agreed:
Licensed to: Tool(s):
Allowed: demo / release / commercial. Platforms:
Not allowed:
Ends: Model and files then: delete / return / keep
Fee or split: Credit as:
Disclosure wording:
Signed (both):
Release checklist
From Chapter 12: Releasing and Disclosure
Go through it for every release. If you cannot tick a line, fix that before you upload.
- The core of the song (lyrics, melody, main performance) is mine, or I know exactly which parts are not.
- Every AI tool used was on a plan that allows commercial use on the day I used it.
- No voice in the track belongs to anyone who has not given written permission.
- No stems separated from a commercial recording, and no samples I have not cleared.
- No lyrics or melody lines reproduced from an existing song, even by a tool.
- My distributor accepts this kind of track under its current AI policy.
- AI disclosure answered per layer: lyrics, composition, recording, artwork.
- Title, artist name and artwork are honest and do not suggest another artist.
- Track length and release pace are normal; no near-duplicate versions.
- Content ID registration only for audio that is substantially my own performance.
- Record sheet filled in and saved with the project files and the final masters.
- Local copies of every AI output and every project file, outside the tool.
Record sheet (one per track)
From Chapter 12: Releasing and Disclosure
Fill it in as you work, not on release day. Keep it with the project files, and keep a copy somewhere else.
| Field | Your notes |
|---|---|
| Title and versionWorking title, final title, version number of the master | |
| DatesFirst idea (with file name of the memo or demo), recording dates, release date | |
| WritersEvery person who wrote lyrics or music, with their share | |
| PerformersEvery person who played or sang, and on which part | |
| AI toolsTool name, plan (free or paid), model version, dates used | |
| PromptsEvery prompt whose output you kept, copied exactly, or the file where they are saved | |
| AI outputs keptFile names of the raw AI outputs, before your edits | |
| What AI didPer layer: lyrics, composition, recording, artwork. One line each | |
| What you didPer layer: what you wrote, played, sang, chose, changed, arranged, mixed | |
| Voices and samplesWhose voice, with a reference to the written permission; every sample and its license | |
| DistributionDistributor, date submitted, AI answers given (copy them exactly) | |
| RegistrationsCopyright registration number and what was claimed and excluded; PRO and Content ID entries |
Disclosure note
From Chapter 12: Releasing and Disclosure
Disclose per layer: lyrics, composition, recording, artwork. Ask of each part: who made the first version?
The book's example, to adapt
Lyrics, vocal melody and chords written by the artist.
Lead vocal, acoustic guitar and bass performed by the artist.
Drum groove and string pad generated with an AI tool, then edited and mixed by the artist.
Mastered with an AI-assisted mastering service.
Cover photo by the artist. No AI voice cloning.
Personal AI policy
From Chapter 13: A Sustainable Practice
Write yours in plain words, on one page. Review it once a year.