AI content rules on writing platforms

The craft questions about AI get most of the attention: is it cheating, will it flatten your voice, does it help. The logistical question arrives later and matters just as much: does the platform where you post allow it, and what must you disclose? The rules differ sharply between sites, they tend to live in help pages rather than terms of service, and they change. Below is what each platform's own pages said when this guide was written, in late September 2026. Quote-check anything that decides something for you, on the platform's own site, in the week you start posting.

Royal Road: allowed, tagged, and answerable

Royal Road's content guidelines engage with AI directly and split it into two tags:

• AI-Assisted: "AI used for editing/proofreading."
• AI-Generated: "The story was generated using an AI tool; the author prompted and directed the process, and edited the result."

The rules around the tags are specific. AI content belongs in chapter pages only: comments, reviews, forum posts and private messages cannot include AI-generated content unless it is a quote. Negligible AI snippets may be tagged merely AI-Assisted, but the guidelines still ask for a warning in the author's notes on the chapter that contains them (the notes being their own craft, see author's notes on serial platforms). Quality is a stated requirement, not a suggestion: "A high quality must be retained. AI-generated content must be moderated and refined by the author to ensure quality, continuity, and readability." The content itself "can not violate any laws and can not violate our rules and guidelines," and its use is at the author's own risk. Even AI translations get a rule: no tag is needed if a human knowledgeable in both languages reviewed the translation; otherwise it must be tagged AI-Generated.

Royal Road's terms of service separately require you to represent that you hold all necessary rights to what you post.

Wattpad: no AI-writing rule on its own pages

Wattpad's content guidelines mention AI exactly once, and not about stories: works are removed if they involve non-consensual AI content, meaning deepfakes of real people. Neither those guidelines nor Wattpad's terms of service, last revised May 2019, restrict, define or require disclosure of AI-written fiction. So the platform's published position, as its own pages read, is silence: the ordinary content rules apply, and there is no AI tag to set. Silence is not encouragement, and it may not last; check the guidelines yourself when you post.

Scribble Hub: the story must be mostly yours

Scribble Hub's content guidelines are the bluntest of the three: "AI Generated Stories - Stories created mostly by AI will be rejected. You can use AI to help create your stories but most of the story should still be written by you." Help is explicitly fine; the line is the share. The page carries a February 2023 revision date, so the wording is not new, but it remains the current published rule.

Inkitt: could not be verified

Inkitt's guideline pages could not be retrieved while this guide was being written, so no claim about Inkitt's position appears here. Treat any secondhand summary, in either direction, as unverified until you have read Inkitt's own pages. That is also the correct default for every platform not listed above.

Three layers of rules, often confused

Writers get into trouble by mixing up three separate rulebooks:

• The AI tool's terms. These govern your relationship with the tool vendor. Anthropic's commercial terms, for instance, state that the customer "owns its Outputs," assign the vendor's rights in those outputs to the customer, and say the vendor "may not train models on Customer Content from Services." Check the terms of the specific tool you use.
• The platform's content rules. These govern what may be posted, where, and with what tags. Owning the output under a tool's terms gives you no pass on a platform's posting rules; the two contracts have nothing to do with each other.
• Copyright law. What you can own, protect and register is a separate question again, decided by law rather than by either set of terms.

Habits that keep you out of trouble

• Read the platform's own guideline page before you post, not a summary of it, and re-read it when platforms revisit these rules. Everything above was current in late September 2026; this area moves.
• Disclose more than required. A line in your description or author's note about how you use AI costs you a few readers at most and buys credibility when the rules tighten. Royal Road's model, a visible tag plus a note on the chapter, is a reasonable template even where no tag exists.
• Tag to the honest side. If the machine drafted the chapter and you edited it, that is the AI-Generated case wherever the distinction exists, not the assisted one.
• Keep the human share real. Every platform that addresses AI content makes the author answerable for its quality, continuity and legality. That is the same deal as choosing writing software at one remove: know what the tool does, and stay the one who signs off.

Where a platform stands on AI is worth weighing before you commit a serial to it, alongside audience and monetization, in the wider serial platform landscape.