Where Does AI Help?
Writers waste hours fighting AI to do things it was never built for. Then they dismiss the whole technology because it failed at impossible tasks. Meanwhile, they ignore the tasks where AI helps.
I sit squarely between the cheerleaders and the doomsayers. Given the right jobs, AI makes a writer faster. Handed the wrong ones, it makes a writer worse, and selling that shortcut to beginners who can’t yet tell which jobs are which is irresponsible. Know the difference and you’ll stop wasting time.
Brainstorming at scale. Need fifty names for a fantasy kingdom? Twenty possible motivations for a betrayal? Fifteen ways a scene could end? AI generates volume. Most options will be mediocre. That’s fine. You need volume to find the gems. Writers staring at blank pages often just need raw material to react against. “None of these work, but option seven has something…” That reaction is valuable. AI gave you something to push against instead of nothing.
Research synthesis. How did Victorian servants address their employers? What does arsenic poisoning look like over time? How do police process a crime scene? AI compresses hours of research into minutes. It explains complex topics in terms fiction writers can use. Always verify critical details, but for directional research and getting oriented in unfamiliar territory, AI saves enormous time.
I’d never put an AI answer about poison or police procedure into a manuscript without checking it. A reader who knows the subject stops trusting everything else in the book the moment one detail turns out to be invented, and the author doesn’t get that reader back.
Pattern recognition in your own work. You can’t see your own tics. After fifty pages, you’re blind to the sentence structures you repeat, the words you overuse, the beats your scenes follow every single time. AI sees them immediately. “You start 40% of your sentences with ‘She.’ Every argument scene ends with someone leaving the room. The word ‘suddenly’ appears 47 times.” Humbling. Also useful.
Continuity checking. Did Marcus have blue eyes in chapter two and brown eyes in chapter nine? Did the sun set twice on the same day? Is the character’s sister named Elena in one scene and Elana in another? Human eyes miss this stuff after multiple read-throughs. AI catches it on the first pass. Feed it your manuscript and ask for continuity errors.
Explaining craft concepts. Don’t understand what “free indirect discourse” means? Confused about the difference between theme and motif? Want to understand three-act structure versus four-act? AI explains craft concepts clearly and can generate examples. It won’t teach you to feel story the way experience does. But for intellectual understanding of technique, it’s a patient, thorough teacher.
Where Does AI Fall Flat?
Judging emotional impact. AI can tell you a scene is “emotionally charged.” It can’t tell you whether readers will feel anything, and that gap is the biggest one on this list. AI analyzes patterns that usually indicate emotion. It doesn’t experience emotion. It can’t tell you whether your romance creates butterflies or your horror scene generates dread. Only human readers can. Never trust AI to tell you if something works, because it has no idea.
Writing in your voice. AI can mimic styles. It can’t replicate the combination of rhythms, word choices, obsessions, and instincts that make your prose yours. AI prose has a smell. Readers are learning to recognize it: sentences that are technically correct but somehow lifeless, competent but never surprising, safe choices everywhere. Your voice is your brand. AI can’t give it to you.
Making creative decisions. Should the mentor die or survive? Should the lovers reunite or stay apart? Should the detective discover the killer was her father? These are author choices. They matter because a human made them. They reflect your vision, your values, your sense of what the story needs. AI will happily make these decisions if you ask. It has no stake in the outcome. Handing your creative choices to something that doesn’t care makes me wince, because those choices were supposed to be the part of the book that’s yours.
Understanding what your story needs. AI knows general principles. It doesn’t know your book. It doesn’t know your protagonist’s arc requires her to fail at this exact moment. It doesn’t know the subplot connects to the theme in ways you never explained. It doesn’t know the “pacing problem” it flagged is a deliberate slow burn you need. AI gives generic advice dressed in precise-sounding language. You have to evaluate whether generic advice applies to your situation. The Revisions Handbook covers how to evaluate feedback and know when to ignore advice that doesn’t fit your vision.
Originality. AI remixes what exists. It can’t create what has never been. Every output is a statistical blend of training data. AI produces competent averages. It doesn’t produce the weird, specific, personal thing that only you could have written. If you want your work to stand out, originality has to come from you. AI can help you execute. It can’t help you innovate. A lot of what these models learned from comes from places like Reddit and Wikipedia, two of the worst sources out there, both edited by people with their own biases. So AI writes at the lowest common denominator, and a lot of what it writes is shit.
What is the hybrid approach to AI writing?
Smart writers use AI for tasks in column A and handle tasks in column B themselves.
Generate twenty brainstorm options with AI. Select and transform them yourself. Let AI check your manuscript for continuity errors. Decide yourself whether to fix them. Ask AI to explain narrative techniques.
Apply them based on your own judgment of what your story needs. Get AI feedback on pacing. Evaluate whether that feedback applies to your book.
The hybrid approach treats AI as a tool that saves time on mechanical tasks while keeping creative authority with the human. You generate more options, catch more errors, and understand more craft. But every decision that matters stays with you.
Which Tasks Do Most Writers Give AI by Mistake?
Writers constantly ask AI to do things in column B and then complain when it fails.
“Write this scene for me.” AI produces something generic. Writer is disappointed. “Tell me if this ending works.” AI says it works fine. Beta readers hate it. “Make my character more interesting.” AI adds surface details that don’t touch the underlying problem.
Readers pay for that mistake. They bought a story and got the machine’s average guess at one, and plenty of them won’t give that author a second try. If you can’t afford an editor or a ghostwriter, writing with AI is fine. Keep the decisions for yourself and it can still be your book.
The failure isn’t AI’s fault. You asked it to do something it can’t do. If you’re frustrated with AI results, check whether you’re asking for brainstorming and analysis (it can help) or judgment and creation (it can’t). Adjust your requests accordingly. Learning to prompt effectively eliminates most of this frustration.
What Should You Ask Before Every AI Prompt?
Before any AI prompt, ask yourself: Am I asking for options or decisions?
If options, proceed. AI can give you raw material, research, patterns, possibilities. If decisions, stop. Make the decision yourself. AI doesn’t care about your story. You do. The Plot Handbook covers frameworks for making structural decisions when you’re stuck.
That single filter will eliminate 90% of AI frustration. Too many writers skip it and then blame the tool for a book they handed away themselves. The machine did what they asked. When the finished book reads like everybody else’s, the writer made that call, and the reader who quits at chapter two is the one who pays for it.
