You use Grammarly, ProWritingAid, or an embedded grammar checker and never think twice. You’ve got a strong opinion about ChatGPT writing your prose. Somewhere in between those two things is a whole stack of tools, decisions, and quiet compromises you’ve never actually sat down and evaluated. That gap between what you use and what you’d admit to using out loud is where most of the confusion lives.
So let’s fix that. Not with a hot take on whether AI is good or bad for publishing. With a framework you can actually use.
The Binary Question Doesn’t Work
“AI or not AI” sounds like a real question. It isn’t. It assumes every tool and every task deserves the same yes-or-no answer, when in reality your grammar checker, your cover artist, your ad copy, and your actual manuscript aren’t the same category of decision at all.
Treat them like one question, and you’ll end up in one of two bad spots. Either you reject tools you’re already using without realizing it, or you wave through something you’d actually be uncomfortable defending to a reader if they asked you directly.
You need more than one bucket. Here’s a framework with three.
Three Questions for Any Tool
Before you decide whether a tool crosses your line, ask it these three questions.
Does it touch my actual prose? Not your workflow. Not your business. The words a reader will read on the page. A tool that fixes a comma splice is doing something very different from a tool that writes the sentence for you.
Does it replace my judgment, or support it? A research assistant that hands you six possible names for a planet and lets you pick and combine two of them is not making the creative call. You are. A tool that outputs a finished paragraph you drop straight into your manuscript is making the call for you, whether you meant it to or not.
Would I be comfortable telling a reader exactly how I used this, in plain language? This is the real test. Not “would they be mad,” but could you say it out loud without flinching or over-explaining? If the honest answer to “how’d you write this scene” makes you uncomfortable, that’s information.
Run any tool through those three questions, and you’ll usually get a clear answer fast.
Applying It Across Your Business
Here’s what that looks like sorted into the categories most authors are actually dealing with.
Editing and grammar tools. Grammarly, ProWritingAid, Word or Google Docs’ built-in checker. These almost never touch your prose in a way that replaces your judgment. You’re still deciding what the sentence says. Low concern for most authors.
Generative writing tools. Anything that produces finished sentences, paragraphs, or scenes meant to go straight into your manuscript. This is where the real line sits for many authors, and it’s worth deciding on purpose instead of by default.
Research and brainstorming. Naming conventions, worldbuilding details, “how would a character experience this,” continuity checks, character bios pulled together for your own review. This is judgment support, not judgment replacement, as long as you’re the one making the final creative call.
Design and marketing tools. Cover design software, ad copy generators, social media schedulers. A lot of these have AI baked in now, whether they advertise it or not. Worth knowing, rarely worth agonizing over.
Business and admin tools. Royalty tracking, email platforms, analytics. AI is often present here too, quietly, and it’s the category authors worry about least and probably don’t need to.
Notice that “does this tool use AI” and “should I be concerned about this tool” are two completely different questions. Almost every category above includes AI somewhere. Only one of them usually needs your actual attention.
Write Your Line Down
Here’s the part most authors skip. Once you’ve run your tools through the framework, write the result down somewhere. Not for anyone else. For you.
A single sentence is enough. Something like: I use AI for research, naming, and continuity tracking, but every word of my prose is mine. That sentence does two things. It settles the question for you so you’re not re-litigating it every time a new tool launches. And it gives you language ready to go if a reader, an agent, or a publisher ever asks directly.
You don’t need a manifesto. You need a line you can say without hesitating.
The Real Point
This isn’t about landing on the “right” answer. Plenty of authors will draw their line in a different place than you do, for reasons that make just as much sense. The point is drawing it on purpose, with a framework, instead of reacting tool by tool as new ones show up.
That’s the difference between having a position and just having opinions. One of those holds up when someone asks you a direct question. The other one doesn’t.
