AI, Authorship & Publishing
What writers actually need to know.
Is using AI to write as risky as you've been told? Some of the fear is justified, and some isn't. Here is what the main rule-makers say, with links so you can check for yourself.
Last reviewed: 11 November 2026
The short version: the rules don't ask "did you use AI?" They ask how you used it, what you contributed, and where you're publishing. Using AI to edit and improve your own writing is treated very differently from having AI write the text. Know which you're doing, and be honest about it.
This page is general information, not legal advice. Rules change and differ between countries and publishers. Always check the current policy of the platform or publisher you're dealing with.
Amazon KDP: assisted versus generated
Amazon Kindle Direct Publishing separates two things:
| AI-assisted | You created the content yourself and used AI to edit, refine, error-check or improve it, or to brainstorm ideas you then wrote up yourself. No disclosure required. |
|---|---|
| AI-generated | Text, images or translations created by an AI tool, even if you edited them substantially afterwards. You must tell KDP when you publish or republish the book. |
Either way, you are responsible for making sure the content follows KDP's guidelines and doesn't infringe anyone's rights.
Source: Amazon KDP Content Guidelines: Artificial intelligence (AI) content
What that means if you use IDEORIX
- You wrote the manuscript and IDEORIX edited it (Bring Your Own Manuscript): that fits KDP's description of AI-assisted work.
- IDEORIX drafted the chapters (Full Pipeline): the text is AI-generated under KDP's definition, however much you directed and revised it, and you should declare it.
IDEORIX supports both ways of working, and keeps an AI-Use Disclosure Log for each project: a factual record of what AI generated and what it assisted. The choice is yours; the disclosure follows from the choice.
Copyright in the United States
In January 2025 the US Copyright Office published its report on whether AI-assisted and AI-generated work can be copyrighted. Its main conclusions:
- Using AI as a tool doesn't cost you copyright. Work you authored remains protected when AI assisted you.
- Material generated purely by AI is not protected. Copyright requires human authorship.
- Prompts alone are not enough, with today's technology, to make you the author of what the AI produces.
- Your own contributions are protectable: text you wrote, your creative selection and arrangement of material, and your creative modifications to AI output.
- Each work is assessed case by case, and you are expected to identify AI-generated parts when registering.
In practice: the more of the expression that is genuinely yours, the stronger your position.
Copyright in the United Kingdom
UK law is unusual. Section 9(3) of the Copyright, Designs and Patents Act 1988 gives protection to "computer-generated" works with no human author. In its Report on Copyright and Artificial Intelligence in March 2026, the government said it would keep that provision under review and may repeal it; no decision or timetable has been announced. Work you write yourself, with or without AI assistance, is protected in the ordinary way.
Sources: Copyright, Designs and Patents Act 1988, s.9 · Lewis Silkin commentary on the March 2026 report
Traditional publishers and agents
Policies are not uniform, and a publishing contract can be stricter than any retailer's rules. Many contracts ask you to warrant that the work is original and that you are its sole author. If you hope to be traditionally published:
- Read the publisher's or agent's AI policy before you submit, and the contract before you sign.
- Be accurate about how you used AI if you're asked.
- Keep your drafts and notes. They are the simplest evidence of your own work.
In October 2026 the Authors Guild published best practices urging publishers to adopt clear written AI policies, and said contracts should not be cancelled over AI use unless the author broke a clear and material term.
Source: Authors Guild, AI Best Practices for Publishers (13 October 2026)
AI detectors
Detectors are not reliable enough to settle the question of who wrote something. The Authors Guild states that "AI detectors—even the best—are not always accurate", that they produce both false positives and false negatives, and that no one should rely on a detector alone to accuse an author or cancel a contract. It notes that plain styles, convention-heavy genres and non-native English writers are flagged more often.
That is why IDEORIX's free Tell Tales tool is deliberately not a detector. It points out phrases and rhythms worth revising and never guesses who wrote them.
Source: Authors Guild, AI Best Practices for Publishers
The ethical objections are real
Many writers object to AI on principle: the models were trained on authors' work, largely without consent or payment, and that dispute is still being fought in courts and legislatures. Those concerns aren't answered by the fact that the tools are useful. Whether and how to use AI is a decision each writer has to make with open eyes, and readers increasingly expect honesty about it.
Sensible practice
- Decide how you're using AI (editing your own words, or drafting) and know which rules follow.
- Disclose where it's required, and don't claim "human-written" for text a machine drafted.
- Make the book yours: direct it, revise it, and stand behind every page.
- Check facts independently, especially in non-fiction.
- Keep your working files.
- Re-check the rules before each book; they are changing quickly.
Where IDEORIX stands: the writer should be in control, not the software and not the AI. IDEORIX works on manuscripts you wrote entirely yourself as well as on AI-assisted drafts, and leaves the decisions, and the responsibility, with you. Read what IDEORIX stands for →