Royal Road's AI Tag Rules, Explained

Two pages hold the rules. The Royal Road AI-text policy is blog post #57, first published in June 2023 and updated since. The "A.I. Content" section of the Content Guidelines (knowledge base article 114) has the short definitions of each tag. The Writathon rules add stricter terms for that event.

This article is a plain-English guide, not legal advice and not Royal Road's own wording. Where the wording matters, it quotes the policy directly. PlotWrite isn't affiliated with Royal Road, and Royal Road's own pages always win.

The policy sorts AI use into three tiers, from no restrictions at all to a required tag.

Spell-checkers and grammar tools that suggest fixes while keeping your voice sit outside the policy. Royal Road says this kind of tool "does not fall under this policy" and can be used "without any sort of restriction or tagging". It goes as far as encouraging the Grammarly plugin.

AI-Assisted covers text that keeps "the author's creativity and structure but it uses the AI's voice and tone". The policy's example is taking a paragraph you've already written and asking an AI to improve it.

Doing that occasionally isn't the trigger. The policy asks authors who do it "excessively", for example using an AI as "your personal proof-reader for every sentence", to tag the story AI-Assisted. The knowledge base definition says the author "has used an AI tool for editing or proofreading", that the text "may use the AI's voice and tone", and that "there may be some negligible amount of snippets generated by AI".

A later update adds a middle path for small amounts of AI-written text. If a novel contains "a negligible amount of AI-generated snippets", you may tag it AI-Assisted instead of AI-Generated, but you should "add a warning in the author notes on the chapter with the AI-Generated content".

AI-Generated is when "the author is prompting the AI to write the content and is merely directing it", or in the policy's other phrasing, "where the author is assisting the AI rather than the AI assisting the author". The knowledge base puts it as: the story "was generated using an AI tool; the author prompted and directed the process, and edited the result."

AI-Generated stories are allowed, with conditions:

• You must tag the story AI-Generated.
• "A high quality must be retained", and low-effort text generation should be avoided.
• AI-generated text can't be used in reviews, comments, forum posts or private messages.
• Even if a story isn't tagged, readers may say in their reviews that they believe it is AI-generated.

One sentence in the policy settles most everyday questions: "While you do not need to mention any 'back-office' assistance by an AI, such as brainstorming or research, you do need to mention it when you copy the text."

In other words, the test is whether AI-written words end up in your chapters. Using AI to brainstorm, research or think through a plot needs no mention. Copying what it wrote into the story does.

Royal Road's own examples are brainstorming and research. Other uses that never put AI words into the text, such as keeping a story bible, checking a chapter for contradictions, summarizing your own chapters for reference, or planning arcs, sit on the same side of the line by the same logic. That's our reading of the policy, not a Royal Road ruling. If you're unsure about a particular use, ask Royal Road support.

• Spell-check and Grammarly-style grammar suggestions: no tag.
• Brainstorming ideas, names or plot options with an AI, then writing it yourself: no mention needed (back-office).
• Research questions answered by an AI: no mention needed (back-office).
• A story bible, continuity checks, summaries and outlines that stay out of the text: back-office, by the same logic.
• Asking an AI to improve paragraphs you wrote: occasional use isn't the target; heavy or every-sentence use means AI-Assisted.
• A negligible number of AI-written snippets in the story: AI-Assisted, plus a warning in that chapter's author notes.
• Chapters an AI drafted from your prompts and outline, which you then edit: AI-Generated.
• An AI translation of your own story: no tag if a bilingual human reviews it; otherwise a tag (see below).

Royal Road treats AI translation on its own terms. If a bilingual human reviews the translation, no tag is needed. If nobody does, it needs a tag, and here the two pages differ: the blog post says AI-Assisted, while the knowledge base says AI-Generated.

If you're publishing an unreviewed AI translation, the safe choice is the stricter tag, or a quick question to Royal Road support.

The policy is open about its reasons. Royal Road doesn't "want to start a witch hunt", automated detectors "can not detect it correctly 100% of the time", and the site doesn't want "rules that we can not reliably enforce". So the system runs on honest tagging, not detection.

In practice that means tagging is on you, and readers can report a story they think should carry a tag. It also means the reaction that matters most usually comes from readers, not moderators. Readers can filter out tags they don't want to see, and some hide AI-tagged stories entirely, so a tag changes who finds your story. That's a reason to tag honestly, not a reason to avoid tagging: readers who discover untagged AI text tend to say so in reviews.

Royal Road's Writathon runs twice a year, from April 1 and from October 31, for five weeks. Its rules say: "The story can't be AI-generated. The story can't be an AI-assisted translation." The idea, in Royal Road's words, is that "you personally need to write 55,555 words".

The next Writathon runs from October 31 to December 5, 2026. Planning and research that never touch the text are outside the words you write, but read the current Writathon page before you enter, because event rules can be stricter than the site's.

Amazon KDP draws a similar line. Its content guidelines ask you to disclose AI-generated content, meaning text, images or translations an AI created, even if you then edited it substantially. AI-assisted content doesn't need disclosing: that covers editing, refining and error-checking your own work, and brainstorming ideas you then develop yourself.

If you plan to stub your serial for Kindle Unlimited later, a workflow that keeps AI on the back-office side tends to stay disclosure-free on both platforms.

Add the AI tag to your fiction the same way you add any other tag. If you're using AI-Assisted for a few snippets, put the warning in the author notes of the chapter that contains them, as the policy asks.

A one-line note is enough: say what was AI-written and why. Readers tend to take a plain note much better than finding out for themselves.

PlotWrite is built around the back-office side: a story brain that remembers your facts by chapter, a canon check that flags contradictions with quotes from your own chapters, and planning tools. None of that writes into your chapters.

It can also draft, with Draft with AI and Auto-write. If you paste that prose into a chapter, the rules above apply, and we'd rather say so plainly than let you find out in a review. We've mapped every PlotWrite feature to the policy on PlotWrite and Royal Road's AI policy.