Using AI doesn’t erase your contribution. But putting your name on the result doesn’t tell us what you actually made.
By Rory Jackson
Imagine you shoot a horror movie in your buddy’s basement. You write the scene, direct the actor and spend forty minutes getting a corpse to sit up without knocking over the laundry detergent. Then you get home and realize you can see the wire.
You use AI to remove it.
Congratulations. You still made a horror movie. Your buddy still wants to know when he can use the basement again.
Now imagine you type “a dead woman sits up in a suburban basement” into a video generator. You make twenty versions, pick the creepiest one and put “Cinematography by Me” in the credits.
You did something. You just didn’t do that.
That’s the distinction I keep coming back to. In Why AI Will Win the War for Special Effects, I argued that filmmakers will use AI when it helps them get a shot they couldn’t otherwise afford. I still think that’s true. But what a tool lets you make and what you can honestly take credit for are two different questions.
There’s No Percentage Where You Stop Being an Artist
I’d love a clean rule. Use AI for less than 20 percent of the work and you’re an artist. Cross 21 percent and we come to your house and take your beret.
Unfortunately, the part that matters most is rarely the part that takes the longest.
Say you’ve painted a horror poster by hand. The face, the light and the expression are yours. You use AI to extend a brick wall on either side so the art fits a wider banner. The new bricks might take up half the image. They still aren’t the reason anyone looks at it.
Now flip it. You spend days building the background and let AI generate the face. It might occupy ten percent of the poster, but it’s the ten percent that sells the whole damn movie. That doesn’t erase the work you did. It does make “I painted this” a lousy description of how the poster came together.
Writing works the same way. An AI tool can suggest cutting three paragraphs you already suspected were dead weight. It can also write the last line of your story, the one that changes what every scene before it means. In both cases you might say, “I only used it for editing.” One of those descriptions is doing a hell of a lot of work.
Picking the Best Version Is a Skill. So Is Making It.
There’s a bad argument on both sides of this thing. One says a person who uses AI contributes nothing. The other says choosing a result is the same as making everything in it.
If you’ve ever watched somebody pick the worst possible take out of six perfectly good ones, you know selection takes judgment. A film editor can change a performance by deciding how long we stay on an actor’s face. That’s creative work. It doesn’t make the editor the actor.
The distinction matters when you prompt an image generator. Maybe you specified the angle, rejected a dozen versions and assembled pieces from several outputs into a poster that finally works. Those choices are real. So are the choices you didn’t make: the expression you didn’t draw, the lighting you didn’t place and the texture you didn’t paint.
“Poster designed by me using AI-generated imagery” tells me what you did. “Hand-painted by me” tells me what you hope I won’t ask about.
The Messy Middle Is Where Most of the Work Happens
A writer starts with a generated draft. The opening sounds like the back of a shampoo bottle, the central argument is wrong and the ending thanks everybody for attending. So the writer pulls it apart. They bring in examples the draft never considered. They change the argument halfway through because they realize they disagree with the premise. They keep two sentences and throw the rest in the trash.
I have no problem calling that person a writer. At some point, rewriting stops being cleanup and becomes the work.
But suppose someone asks AI to write a story about a priest who recognizes the demon possessing a child. The generator decides how he knows it, what he did the last time they met and why he lies about it to the family. The person changes the priest’s name and fixes a typo. They supplied the setup. The machine supplied the story that makes the setup worth reading.
Hours aren’t a reliable measure, either. A writer can decide in thirty seconds that the priest is lying to protect the demon, then spend three weeks making that decision work. Those thirty seconds might matter more than the entire cleanup pass. Ask who made the choices you’d miss if somebody took them out. That’ll tell you more than a stopwatch.
Yes, That Applies to This Article
This piece started with a question I wanted to ask. I used AI to help develop the argument and draft the prose. That’s more than spell-check, and it would be bullshit to pretend otherwise.
Putting my name on it means I’m responsible for the argument. It means deciding what I actually agree with, cutting what doesn’t sound right and being able to defend the point without asking a chatbot what I meant. It does not mean I get to claim I wrote every sentence without help.
There’s also a difference between standing behind an argument and doing enough work to earn a byline. If all I did was read a generated draft, think “sounds about right” and hit Publish, that’s approval. It isn’t the same as writing. This article doesn’t get an exception just because it’s the one making the rule.
Take Credit for the Work You Did
There’s another argument worth having about where the material used to train these tools came from and whether the people who made it had a say. An honest credit line doesn’t settle that question. It does settle the smaller one right in front of us: what are you asking the audience to believe about your contribution?
If you wrote the story and used AI to spot pacing problems, say that when it matters. If you built the poster around an image you generated, call it design and take credit for the design. If the machine supplied the argument, the examples and the prose, don’t sell me a story about your lonely struggle with the blank page.
And if you made the corpse sit up and AI took out the wire, it’s still your damn corpse.
