Can You Copyright an AI-Written Book in the UK? (And Can KDP Reject It?)

Last reviewed by Robert Prime — July 2026

Short answer: it depends on how much of the book is genuinely your own work — and the UK and US answer it differently. If you wrote and shaped the book yourself and used AI as a tool (drafting help, editing, brainstorming), you have ordinary copyright and you own it, the same as any other author. If the book is wholly machine-generated with no meaningful human authorship, protection gets shaky fast: in the UK it falls under a peculiar, contested provision for "computer-generated works"; in the US, purely AI-generated material cannot be copyrighted at all. And copyright is a completely separate question from whether Amazon KDP will let you publish — KDP doesn't ban AI books, it just makes you disclose them.

This is general information, not legal advice — confirm your own situation with a qualified adviser (the Society of Authors offers members legal guidance).

TL;DR — Key Takeaways

• AI-assisted books are yours. You wrote and shaped it; AI helped. Standard UK copyright applies automatically — no registration needed, protected for your life plus 70 years. This covers the vast majority of "I wrote it with AI" books.
• Wholly AI-generated books are the grey zone. In the UK, CDPA 1988 s9(3) grants "computer-generated works" to "the person by whom the arrangements necessary for the creation of the work are undertaken" — protected 50 years, not life + 70. But this is uncertain law (see below).
• The US is stricter. Purely AI-generated content cannot be copyrighted — human authorship is a hard legal requirement (Thaler v Perlmutter, 2025). Your own human contributions still can be.
• The UK rule may not survive. The Government's March 2026 Report on Copyright and Artificial Intelligence floated repealing s9(3), then concluded more evidence is needed. Treat it as unsettled.
• KDP is a separate question. Amazon does not reject books for being AI-made. It requires you to disclose AI-generated text, images or translations. Rejections come from non-disclosure, quality/IP breaches or metadata issues — not AI use itself.
• The more human authorship, the stronger your rights in both countries. Keep your drafts and edits as evidence.

The nuance: "written with AI" almost always means AI-assisted

Most authors who say they "wrote a book with AI" did something like this: outlined it themselves, prompted a tool for a rough draft, then rewrote, restructured, fact-checked, edited and polished. That is AI-assisted work, and it carries real human authorship — your selection, arrangement, editing and original writing.

That distinction is everything. Copyright protects human creativity. Where a human has genuinely shaped the work, you have an author and you have copyright. Where the human did almost nothing — typed a one-line prompt, took the raw output, hit publish — there may be no human author at all, and that is where protection thins out or disappears.

So the honest answer to "can I copyright an AI-written book?" is usually yes — because it's really an AI-assisted book, and you're the author. The distinction is decisive, and the problems begin only at the fully-automated end of the spectrum. If you're weighing up which tools keep you safely on the "assisted" side, our guide to AI writing tools for indie authors covers the workflows that preserve your authorship.

Can you copyright an AI-written book? The UK vs US answer

The two markets most UK authors publish into take genuinely different positions. This table is the core of it:

The UK position, in detail

UK copyright is automatic and requires no registration — the moment you fix an original work, you own it (this is covered fully in our how to copyright a book in the UK guide). The only question AI raises is whether an "original literary work" with a human author actually exists.

Two paths:

1. AI-assisted — you contributed real skill and judgement (writing, editing, structuring, selecting). This is an ordinary literary work; you are the author; full protection.
2. Wholly computer-generated — section 178 of the CDPA 1988 defines this as work "generated by computer in circumstances such that there is no human author." For these, s9(3) says the author "shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken," and s12(7) sets protection at 50 years from the end of the year of creation.

Here's the catch, and why you should not lean on s9(3):

• It is one of very few such provisions in the world and has barely been tested in court.
• It sits awkwardly with the modern originality standard UK courts apply — that a work must be its author's own intellectual creation, which implies human creativity.
• Its future is openly in doubt. In the Government and UK Intellectual Property Office (IPO) Report on Copyright and Artificial Intelligence (published 18 March 2026, following the December 2024–February 2025 consultation), ministers signalled they are minded to remove protection for wholly computer-generated works entirely — the stated direction of travel — but stopped short of legislating, saying more evidence is needed first.

Bottom line: the law here is genuinely unsettled, and the current trend points towards less protection, not more — do not treat s9(3) as a reliable safety net.

The US position, in detail

The US line is clearer and stricter: human authorship is required as a matter of law. In Thaler v Perlmutter, the DC Circuit affirmed in March 2025 that a work generated autonomously by AI, with no human author, cannot be registered — and the Supreme Court declined to hear the case in March 2026, leaving that ruling in place.

That does not ban copyright for books made with AI. The US Copyright Office's guidance (its March 2023 registration guidance and its January 2025 Copyrightability report) is consistent: you can copyright your own contributions — the text you wrote, and creative selection, arrangement or modification of AI output that reaches the originality standard. When you register, you must disclose the AI-generated material and disclaim it, describing what you the human added. What you cannot do is claim the raw AI output itself — and the Office has said that prompts alone don't give a user enough control over the result to count as authorship of that output.

What this means for KDP (and "can Amazon reject it?")

This is where authors get the two issues tangled. Copyright ownership and Amazon KDP's policy are different things. Copyright is set by law (above). KDP is a platform with its own contract terms.

The important facts:

• KDP does not ban AI-made books. You can publish a book written with — or even entirely by — AI. Amazon does not reject a book simply for being AI-generated.
• KDP requires disclosure. During publishing, KDP asks whether your book contains AI-generated content (text, images including the cover, or translations). AI-assisted work — where you created the content and AI helped you edit or refine — needs no disclosure. We break the exact line down in our KDP AI content disclosure guide. The disclosure is internal; readers never see it on your product page.
• Watch the overlap — the two "assisted" tests aren't the same. This is where authors trip. A book can be AI-assisted for copyright (you rewrote and shaped an AI draft, so you own it) yet still count as AI-generated for KDP's disclosure rule — because Amazon treats text an AI tool produced as AI-generated even if you edited it heavily afterwards. If an AI tool wrote the underlying draft, disclose it to KDP, regardless of how much you own the copyright. The copyright test and the disclosure test draw the line in different places.
• What actually gets books rejected or removed is non-disclosure, breaching content or quality guidelines, intellectual-property problems, or metadata issues — not the use of AI in itself. Amazon also caps how many new titles you can create per day (a small daily limit) to slow low-quality content farming.
• You still warrant that you hold the rights. KDP requires you to confirm you own or control the necessary rights. If your book is pure AI output with thin or no copyright, you can still publish it — but you have little legal protection if someone copies it, and you must not misrepresent what you own.

So: can KDP reject an AI-written book? Not for being AI. It can — and does — remove books for lying about AI content, poor quality, or IP breaches. The safe path is to disclose honestly and make the book genuinely good.

The practical takeaways

• Do real human work and you keep real rights. The more you write, edit, structure and shape, the stronger your copyright — in both the UK and US. AI-assisted is the safe, protected default.
• Don't publish raw, unedited AI output and expect to own it. In the US you may own nothing; in the UK you're relying on a provision that may not survive.
• Keep your evidence. Save dated drafts, outlines and edit histories. They prove human authorship if it's ever questioned — and help with the proof-of-creation issue that matters for any copyright dispute.
• Disclose AI content to KDP honestly. It's free, it's private, and getting caught misrepresenting it is what costs you the book — or the account.
• Check your AI tool's terms. Major tools currently assign output rights to the user, but that contract can't create copyright where none legally exists — it only affects what the vendor claims.
• When in doubt on anything high-stakes, get advice. This guide is general information, not legal advice.

Owning your book is only half the battle

Copyright decides who owns your book. Discoverability decides whether anyone finds it — and increasingly that's ChatGPT and the other AI assistants doing the recommending, not just Google. This new layer, sometimes called AI book discovery or answer-engine optimisation, decides whether AI names your book when a reader asks what to read. Our £29.99 AI Discovery Audit tests ChatGPT, Claude, Perplexity, Gemini and Amazon Rufus on your title and hands you the exact fix list, and the AI Discoverability for Authors course walks you through the whole approach. No pressure — the guides above answer the copyright question on their own.

Frequently Asked Questions

Can I copyright a book I wrote with the help of AI in the UK?

Yes. If you genuinely wrote and shaped the book and used AI as an assisting tool, it's an ordinary literary work with you as the author. UK copyright is automatic on creation — no registration — and lasts your lifetime plus 70 years. The only doubtful case is a book that is wholly machine-generated with no meaningful human input.

Who owns the copyright in a fully AI-generated book in the UK?

Under s9(3) of the Copyright, Designs and Patents Act 1988, a "computer-generated work" with no human author is attributed to "the person by whom the arrangements necessary for the creation of the work are undertaken," for 50 years. Exactly who that is — the person operating the tool, or arguably the tool's developer — is itself unresolved. And the whole provision is contested, rarely tested in court, and the Government signalled in its March 2026 report that it is minded to remove it, so don't rely on it for strong protection.

Can you copyright AI-generated content in the US?

No — not the AI-generated part. US law requires a human author (confirmed in Thaler v Perlmutter, 2025). You can copyright your own human contributions — text you wrote, and creative selection, arrangement or modification of AI output — but you must disclaim the AI-generated material when you register with the US Copyright Office.

Will Amazon KDP reject my book for being written by AI?

No. KDP does not ban AI books. It requires you to disclose AI-generated content (text, images or translations) when you publish; AI-assisted work needs no disclosure. Books get removed for failing to disclose, for quality or content-policy breaches, or for IP problems — not for using AI. See our KDP AI content disclosure guide.

Is UK or US AI-copyright law more author-friendly?

For fully AI-generated work, the UK is nominally more generous — s9(3) offers a (contested) route to 50 years of protection, whereas the US grants none. But for the far more common AI-assisted book, both protect your human authorship, and the US position is clearer and more settled. In practice the difference only bites at the fully-automated end.

Does disclosing AI use to KDP affect my copyright?

No. The KDP disclosure is a platform requirement for Amazon's internal use; it doesn't change who owns the copyright, which is decided by law. The two are entirely separate — you can owe a disclosure and still own full copyright, or owe no disclosure and still have thin rights.

About the Author

Robert Prime is a self-published author, veteran eCommerce strategist, and the founder of publishing.co.uk. With over 25 years in eCommerce and digital publishing — including founding the Amazon strategy consultancy MrPrime — he brings a practical, numbers-first perspective to self-publishing. After navigating the formatting and marketing of his own book, Google. Panic. Repeat., he built publishing.co.uk to help UK authors avoid the same pitfalls. He co-owns the LoveReading network.

This article is general information, not legal advice. AI and copyright law is developing quickly and differs by country — confirm your own situation with a qualified legal adviser before relying on it.