Cookie Settings

    We use cookies to analyse website usage and improve the experience. You can adjust your settings at any time.

    More information in our Privacy Policy.

    Back to overview
    Industry

    AI Disclosure for Authors: What UK and US Rules Mean for Your Books in 2026

    August 31, 20268 min

    No EU AI Act in the UK or US โ€” but Amazon KDP has mandatory rules, and new legislation is coming. A practical guide to what you actually need to do.

    The Regulatory Landscape Is Fragmented

    If you are an author in the United Kingdom or the United States, here is the headline: there is no equivalent of the EU AI Act in your country. No single, comprehensive law forces you to label AI-assisted books. But that does not mean there are no rules.

    The binding constraint for most indie authors is not legislation -- it is platform policy. Amazon KDP has required AI disclosure since September 2023, and that requirement carries real consequences: book removal and account suspension.

    Meanwhile, lawmakers on both sides of the Atlantic are working on proposals. The landscape is moving, and it pays to understand where things stand right now.

    For comparison, the EU's AI Act (Regulation 2024/1689) includes Article 50 transparency obligations that took effect August 2, 2026. If you sell books in EU markets, those rules may touch you too -- we cover that further below.

    United Kingdom: Pro-Innovation, Voluntary Approach

    The UK government has consistently favored a pro-innovation stance on AI regulation. Unlike the EU, there is no overarching AI act, and no mandatory labeling requirement for AI-generated text.

    Where Things Stand

    • The UK's March 2026 government report established a taskforce to propose best practice guidelines for AI content labeling. An interim report is expected in autumn 2026 -- but these will be recommendations, not law.
    • The UK AI Bill, which would give statutory footing to some AI governance principles, has been delayed until at least summer 2026. Its scope remains uncertain, and labeling of creative content is not a confirmed focus.
    • During the 2025-2026 public consultation, respondents broadly agreed on one distinction: fully AI-generated works should carry labels, but AI-assisted works -- where a human directs and revises -- should not require mandatory disclosure.

    Current Status

    As of mid-2026, AI content labeling for books in the UK is entirely voluntary. There is no legal penalty for publishing an AI-assisted novel without a disclosure notice. That said, platform rules still apply -- and Amazon is the platform that matters most.

    United States: No Federal Law, State Patchwork

    The US has no federal AI labeling law for text or books. But that does not mean the regulatory landscape is empty.

    Federal Level

    • The FTC (Federal Trade Commission) uses its existing Section 5 authority against deceptive practices. Its "Operation AI Comply" initiative has targeted misleading AI claims -- but there is no specific mandate requiring book authors to label AI usage.
    • Several bills have been introduced in Congress but none has passed. The REAL Act (pending as of 2026) would require labels on AI-generated content, but includes an editorial responsibility exception -- if a human exercised substantial creative control, the labeling threshold may not apply.

    State Level

    • California SB-942 (effective August 2026) is the most prominent state law, but it covers only image, video, and audio content. Text is explicitly excluded from its scope.
    • Other states have introduced or passed laws targeting AI-generated deepfakes, election content, and synthetic media -- but none specifically covers books or written fiction.

    The result is a patchwork where no single rule applies to book authors, but the direction of travel is clearly toward more transparency.

    Amazon KDP: The Rule That Actually Matters

    For self-publishing authors, Amazon's policy is the most practically binding requirement -- more consequential than any current UK or US law. It has been mandatory since September 2023.

    AI-Generated vs. AI-Assisted

    Amazon draws a clear line:

    AI-Generated (must disclose):

    • Text, images, or translations created by an AI tool, even if you edited the output afterward
    • Cover art generated by AI tools (Midjourney, DALL-E, etc.)
    • Machine translations, even if reviewed

    AI-Assisted (no disclosure required):

    • Using AI for brainstorming, outlining, or ideation
    • Grammar and spell-check tools
    • Research assistance
    • AI-powered editing suggestions that you accept or reject

    Key Details

    • The disclosure is internal to Amazon -- it is part of the publishing workflow and is not visible to readers on the product page.
    • Penalties for non-disclosure include book removal and account suspension. Amazon has enforced this, particularly against mass-produced AI content.
    • Amazon also implemented a 3-book daily upload limit alongside its AI policy, targeting content farms.

    Practical KDP Disclosure for SYMBAN Users

    When publishing through KDP, select "AI-assisted" and use wording along these lines in the disclosure field:

    "AI tools were used to assist with drafting and revision. The author directed all creative decisions including plot, characters, style, and narrative structure. All content was reviewed, revised, and approved by the author."

    This accurately reflects the SYMBAN workflow: you set the concept, the pipeline generates and revises drafts, and you approve or modify every chapter.

    How SYMBAN Makes Compliance Easy

    Regardless of which jurisdiction's rules you are following, SYMBAN's architecture supports transparent, documentable AI usage.

    Transparent Pipeline

    The 5-pass pipeline documents every step automatically. Each chapter records which passes ran, what the AI contributed, and what you changed. This creates a natural audit trail you can present to any platform or publisher.

    Clear Human-in-the-Loop

    • Concept and direction: You define the premise, characters, arcs, and style before a single word is generated.
    • Revision control: Automated quality checks flag issues, but you decide what gets changed.
    • Approval gates: No chapter is finalized without your explicit sign-off.

    Metadata Export

    SYMBAN stores pipeline metadata for every chapter. On export, this data can serve as documentation for KDP disclosure, publisher inquiries, or any future regulatory requirement.

    What About the EU AI Act? (For UK/US Authors Selling in Europe)

    If you distribute books in EU markets -- and most KDP authors do, whether they realize it or not -- the EU AI Act's Article 50 transparency obligations are worth understanding.

    The Basics

    Since August 2, 2026, deployers of AI systems in the EU must disclose when content is AI-generated. But the regulation includes important exceptions:

    • Fiction and creative works: The obligations are reduced for content that is "manifestly artistic" in nature. A novel clearly falls into this category.
    • Editorial responsibility: When a human exercises editorial control over the content -- selecting, revising, approving -- the transparency threshold is lower.
    • Practical scope: The regulation primarily targets AI providers (the companies building the models), not individual authors using those tools.

    What This Means for You

    For fiction authors who actively direct and revise their work, the practical impact of the EU AI Act is minimal. A brief transparency note on the copyright page is good practice but likely exceeds what the regulation strictly requires for editorially controlled fiction.

    Common Misconceptions

    "I cannot use AI for my books anymore." Wrong. No UK, US, or EU regulation prohibits AI-assisted writing. The trend is toward transparency, not prohibition.

    "I must label every AI-generated word." No. A general disclosure -- either on the copyright page or through the platform's upload process -- is sufficient. No jurisdiction requires word-level attribution.

    "These rules only apply to big companies." Amazon KDP's AI disclosure rules apply to every publisher on the platform, from major houses to first-time indie authors. Size does not matter.

    "If I edit the AI output heavily, I do not need to disclose." Under Amazon's policy: wrong. If AI created the initial draft, it is "AI-generated" regardless of how much you edited. The distinction is about origin, not final quality. The correct classification for SYMBAN users is typically "AI-assisted," because the human directs the creative process from concept through approval.

    Checklist: What You Should Do Now

    • Complete KDP AI disclosure during upload -- select the appropriate category and describe your process
    • Use "AI-assisted" classification if you actively direct the creative process and revise outputs (this fits the SYMBAN workflow)
    • Keep documentation of your creative process -- concept notes, revision history, approval decisions
    • Stay informed on the UK AI Bill (expected summer 2026+) and US federal proposals (REAL Act and successors)
    • Add a copyright page note for transparency, even where not legally required -- it builds reader trust
    • Review your KDP backlist -- ensure previously published AI-assisted titles have proper disclosure

    For a step-by-step walkthrough of publishing your SYMBAN project on Amazon, see the KDP publishing guide.

    Related Articles