The Difference Between Using AI to Write and Using AI to Code.

The publishing world is dealing with a crisis of ensuring that the authors they publish do not use AI in their work. You would hope that the author’s word on whether they use AI to write the story, AI to edit spelling and grammar, or use AI at all is trustworthy. But there have been a few cases where the author said they did not use AI, but did, such as the story “The Serpent in the Grove” which won an award and was published by Granta, a respected British literary journal.

And then there are cases where the author said they did not use AI in their writing, but their editor used AI, such as Shy Girl published by Hachette Book Group.

The publishing world is now demanding AI disclosure clauses in contracts. Some journals have banned AI-assisted submissions entirely. There are detection tools being deployed, though none of them are foolproof. The whole thing has turned into a cat and mouse between writers trying to hide AI usage and publishers trying to catch it.

But when it comes to coding, there’s no scrutiny for using AI. Businesses that were built using AI have no issues with saying they coded with AI.

In fact, they brag about it. Startups pitch investors on how efficiently their small team shipped a product because they leaned on AI tools like Claude Code.

Even those who know how to code use AI to simplify and speed up their production.

But if a writer who knows how to write uses AI to speed up their production, it will cancel their career.

That double standard sounds unfair on the surface. But the difference is quite obvious. When it comes to coding, the code is not the final product. The product is what the code is building. When it comes to writing, the words that AI writes are the final product.

A reader buys a novel to experience a specific voice. They want the texture of how that particular human mind processes the world. Nobody buys a software subscription to appreciate the elegance of the underlying code.

Besides, AI written text is bad. It’s bad because it’s repetitive. Like two plus two will always equal four. AI writing always results in the same rhythm. Human writers vary in their sentence length. AI usually writes the same medium-long sentences. It also loves to use dashes. Most readers can detect if something was written by AI because the evidence is always the same.

But nobody can determine if a website or app was created using AI to code or if the coding was done entirely by a human. The end user doesn’t care. People only care if apps and programs are useful.

In fact, the use of AI should enhance the production of someone who already knows coding. A skilled developer using AI still makes the architectural decisions. They still debug, review, and own the output. The AI is a tool, like a calculator is a tool for a mathematician.

But if Stephen King used AI to write, it wouldn’t enhance his prose at all. It would just make him look stupid.

Because what makes Stephen King worth reading isn’t the plot. It’s how he writes a sentence. It’s the specific way his brain narrates fear. You can’t automate that. And the moment you try, you’ve already lost the only thing that made you worth reading in the first place.