How do I handle the "AI Disclosure" on Amazon?

This is one of the most anxiety-inducing questions in indie publishing right now, and it does not need to be. Let me cut through the noise.

As of January 2026, Amazon updated their AI disclosure policy alongside several other major changes. When you upload a book through KDP, you are asked whether your content is AI-generated or AI-assisted. What those categories actually mean in practice:

AI-generated means an AI produced the content and you did not substantially edit, rewrite, or create it yourself. This includes text generated by ChatGPT, Claude, or similar tools that you published with minimal human revision. It also includes cover images fully generated by Midjourney, DALL-E, or Stable Diffusion without significant human design work layered on top.

AI-assisted means you used AI as a tool in your creative process, but a human did the actual creating. This includes using AI for brainstorming, research, grammar checking, proofreading, plotting assistance, or generating ideas that you then wrote yourself. It also includes cover designs where a human designer used AI-generated elements as part of a larger, human-directed design.

The practical line: if you wrote the book and used AI to check your grammar, help you brainstorm, or clean up your prose, that is AI-assisted. Check that box, move on, and do not lose sleep over it. If AI wrote the majority of the text and you just cleaned it up, that is AI-generated, and you need to disclose that.

The January 2026 changes also brought several other shifts worth knowing about. Amazon now offers DRM-free EPUB and PDF downloads for ebooks, which means readers can get your book in open formats. They also tightened requirements around alt text for images in ebooks (making accessibility a ranking factor, not just a nice-to-have). And they improved their "Look Inside" feature with longer read samples, which means your opening pages matter more than ever.

What does this mean for you practically? First, document your AI usage as you work. Keep a simple note: "Used Claude for proofreading chapters 3-7. Used Midjourney for initial cover concept, then had designer refine." If Amazon ever asks (and they rarely do for AI-assisted work), you have a clear answer.

Second, do not over-disclose out of fear. Using Grammarly is AI-assisted. Using ProWritingAid is AI-assisted. Using spell check is AI-assisted. You do not need to flag any of these as if you committed a crime. The disclosure is about transparency, not confession.

Third, keep your opening pages strong. With longer read samples now visible to browsers, the first chapter of your book is doing more sales work than ever. Make those pages count. No throat-clearing, no slow warm-ups, no generic introductions. Get to the value immediately.

The authors who are getting flagged or penalized are the ones uploading obviously AI-generated content at scale with zero human oversight. If you are reading this book and care enough to ask this question, you are almost certainly not in that category.

One final point: the disclosure rules will continue shifting. Amazon has already updated their policy twice since initially introducing AI disclosure requirements. Expect further refinements as the technology evolves and as legal frameworks (see Q62 in the next chapter) solidify. The safest strategy is also the simplest: be honest, be transparent, and create genuine value. Platforms reward authors who treat disclosure as a trust-building exercise, not a legal minefield to navigate around.

---