Authors

Is Using AI to Write Cheating? The Answer Most Writers Aren't Expecting.

May 2026

By Mark Reeves — author of all 47 AI Field Guides. Written and tested in a working business.

AI-generated portrait. Pen name. Real work.

48% of authors don't use AI at all. 84% of those cite ethics (Authors A.I., September 2025). That's a serious moral position — and it deserves a serious answer.

You've finished your second novel. Your editor mentioned AI in passing. Now you're wondering whether any of the tools you tried count as cheating — and whether you need to disclose.

The people who aren't using AI because they think it's wrong are not being naive. They're responding rationally to a genuine ethical ambiguity. The fact that 48% of authors are sitting this out, with 84% of them citing ethics as the reason (Authors A.I., September 2025), is not a sign that they're behind the times. It's a sign that the ethics question is real and hasn't been resolved.

This isn't a case for AI. It's a genuine attempt to think through the question. (The ethics question is the right starting point — but it opens into two practical questions most writers hit immediately. There's a sorting exercise at the end that takes five minutes and clarifies exactly where you stand.)

"Using AI to Write" Is Not One Thing

The ethical question collapses immediately when you realize it's actually dozens of different questions dressed in the same clothes.

"Using AI to write" covers all of this: having an AI generate your entire manuscript and publishing it under your name. Asking an AI to check whether your timeline is consistent across 300 pages. Using an AI to brainstorm five possible endings. Running your prose through an AI grammar checker. Asking an AI to describe what a specific street in Lisbon looked like in 1967 so you can fact-check your scene.

These are not the same ethical category. They don't involve the same amount of creative labor. They don't have the same relationship to your voice, your judgment, or your authorship.

The question "is using AI to write cheating?" only makes sense if you specify which of these activities you mean.

The Activities Most Writers Treat as Settled

In practice, most working writers — including many who describe themselves as skeptical of AI — treat certain uses as uncontroversially acceptable.

Research assistance. Continuity checking. Grammar and readability editing. Brainstorming alternative plot directions. Outlining structural options. None of these produce prose that appears in your book. All of them support your judgment and creative decisions. They're tools that help you do your job better, in the same category as a thesaurus, a style guide, or a trusted beta reader.

If you've ever used spell-check, you've accepted that a piece of software can improve your text without you writing every character yourself. The ethics of grammar assistance are genuinely settled. The ethics of AI-assisted brainstorming are only slightly more complex. The ethics of generated prose are much harder.

The Activities That Remain Genuinely Contested

Generating prose that you publish under your name without disclosure is where the real ethical debate lives.

The strongest argument against it: authorship is a promise. When a reader picks up a novel with your name on it, they're assuming that your mind, your experience, your particular way of seeing the world shaped those sentences. If that's not true — if the sentences were generated by a model trained on other people's writing — then the reader is being misled about what they're buying.

That's not a trivial concern. It's actually the heart of why so many serious writers are uncomfortable with this.

The strongest argument for it: every writer is shaped by the writing they've read. Voice is influence all the way down. A novelist who learned their craft by reading Toni Morrison and Hilary Mantel is producing work shaped by those writers. AI is an extension of that influence, not a categorically different thing.

I don't think either argument fully wins. The question is live and will probably stay live for years.

What the Law Says (Which Is Not the Same as Ethics)

Regardless of where you land ethically, there are some facts worth knowing — because they affect you even if you've decided never to use AI.

The U.S. Copyright Office has ruled that AI-generated text is not copyrightable. If an AI wrote it, no one owns it. That means any book with substantial AI-generated content has a copyright problem, not just an ethical one.

Amazon KDP now requires AI disclosure for books sold through its platform. The disclosure system is imperfect and largely self-reported, but the policy exists. If you're publishing on KDP, you're expected to flag AI-assisted content.

These are facts, not arguments. They apply to you whether you use AI or not, because you're operating in an industry where your competitors are making decisions that affect how readers, publishers, and platforms think about authorship.

The Question Underneath the Question

This debate is actually about something deeper for most writers.

The real question isn't "is this cheating?" The real question is: "Am I still a real writer if I use AI?"

That's worth taking seriously. Writing isn't just output. For most serious writers, it's identity. It's how they make sense of the world. The fear isn't just about getting caught or violating a rule. It's about losing something — the discipline, the struggle, the craft — that feels constitutive of who they are.

Here's my honest answer to that: tools don't define the work. Judgment and voice do.

A writer who uses AI to check their plot structure and then rewrites every sentence is exercising the same creative judgment they always have. A writer who publishes whatever the AI produces is doing something different — and most readers can tell, because the work tends to lack the specific gravity that comes from a mind grappling seriously with a story.

The question worth asking isn't "did I use AI?" It's "did I make the hard decisions?" Voice, structure, the willingness to tear apart a chapter that isn't working — that's the job. AI can help you get there. It can't do that for you.

Here's the sorting exercise. Take the last piece of writing where you used AI. For each section, ask: did I make the structural decision, or did AI? Did I make the voice decision, or did AI? Where the answer is "AI did," that's where to apply judgment. It takes five minutes and gives you a clear read on where you stand.

Two things this article doesn't settle: which specific tool actually preserves your voice when you're doing long, sustained work (Claude and ChatGPT behave very differently here), and what you concretely need to disclose to agents and publishers now that some contracts include AI clauses. Knowing where you stand ethically is the prerequisite — but those two questions come right after.

The ethical question is important, but it opens into a set of more practical questions that writers run into immediately. This article addresses the "is it cheating" question. It doesn't address three things that come right after:

Claude vs. ChatGPT for writing — which one actually preserves voice. The tools behave very differently when voice-sensitivity matters. The guide covers this in depth, with specific prompting approaches that change what you get.

How to use AI for outlining and plotting without it taking over. There's a pattern where AI-assisted structure produces technically sound plots with no emotional stakes. The guide covers how to use these tools as a pressure-tester for your ideas, not a replacement for them.

Publisher and agent AI disclosure policies — what you actually need to tell them. This is moving quickly. Some agents now ask about AI use in query letters. Some publishers have contract language covering it. The guide covers what's actually in place as of 2026.

Related reading

• The best AI tools for fiction writers
• How to write better with Claude without sounding like a robot

Mark Reeves is a pen name. AI Field Guide publishes role-specific, practical guides for using AI tools in real work.