The Turing Test of Prose

I built my career writing software. I built bridges between business objectives and the systems and operations that support them. I’ve written elegant, maintainable code, and I’ve written trash. You see, I allow myself to fail, sometimes spectacularly, and I’ve learned to do it right.

One thing stands out from the millions of lines of code I’ve reviewed and maintained: you can tell when someone cared. From the names they used that were well thought out, that clarified rather than obscured, to design patterns that guided you forward instead of fighting you at every turn. You can see it in the edge cases they handled before they became catastrophes, tests that proved the intent wasn’t accidental, that someone had imagined the failure and chosen not to look away.

And you could tell just as clearly when they hadn’t. Logic that worked, but only along the happy path. Patches layered over assumptions nobody had documented. Entire systems that functioned, technically, while resisting understanding at every level. Care left artifacts. So did its absence.

Prose is no different.

Readers won’t tell you they’re noticing structure. They won’t say: the sentence rhythm shifted here and I felt it, the delay before that revelation was exactly right, something was withheld and now I’m leaning in. Often, they don’t analyze. They experience.
It doesn’t matter what genre the prose is written for. We consume stories to feel hope, to imagine ourselves stronger, more capable, and braver than what we see in the mirror.

The author’s voice isn’t decoration, it’s selection, and commitment. It’s what connects readers to us, to our characters, and ultimately to themselves. For our part, as composers of prose, choices are the difference between capturing imaginations and reading weather reports. What gets noticed. What gets ignored. What is stated directly and what is allowed to remain implied, to drift, to land later or never, to become the reader’s discovery rather than the writer’s explanation. A strong voice narrows the world. It says: this and not that. This angle, this light, this moment, rather than the ten moments surrounding it.

I am amazed by what Artificial Intelligence can do. New capabilities come to light every day, every hour. Truly, the latest accomplishments are stunning. And yet.

Generated texts tend to broaden. Machines want to include everything, explain, smooth, and complete it. The generated text is coherent. It is readable. It is, in the way of all things, optimized for general approval, anonymous. No one is standing behind those words because standing behind words requires having chosen them over other words you could have chosen, meaning it, and being willing to be wrong about that choice.

AI tends toward even distribution. Clarity early. Resolution quickly. Confusion is avoided where possible. But clarity too early is a kind of loss. Meaning depends on timing, and without timing, prose flattens into information, which is a different thing entirely from experience.

Great writing is not just what is said. Writers who care think in sequence: what does the reader know now, what happens if I delay this, what gains weight if it arrives later rather than sooner. Structure creates tension the way a held breath does. The release matters because the holding came first.

At the sentence level, the difference is unmistakable. Writers who care break rhythm when it matters. They choose specificity over safety, the word that is exactly right over the word that will do. They let a sentence end sooner than expected and let the silence after it carry weight.

What is left out matters as much as what is included. More, sometimes. Omission creates participation. It creates interpretation. It signals trust. It says: I believe you can hold this gap without needing me to fill it. Generated prose fills gaps because, from a statistical perspective, gaps look incomplete. But those gaps are where meaning lives. When everything is said, nothing is discovered. The reader becomes a passenger instead of a collaborator, and something essential drains out of the experience.

The deeper risk of AI-generated text isn’t about artificiality. It’s scaled indifference. Writing produced without anyone standing behind the choices. Over time, that becomes the baseline, and we are worse for it in ways that are difficult to name and easy to feel.

The presence of a tool does not remove responsibility. It clarifies it. Because now the distinction is unavoidable. The tool produces possibilities. The author chooses what is worth keeping. That choice is where care lives.

Readers notice the difference. Not consciously. But they feel the accumulation of choices made on their behalf, choices that didn’t have to be made, decisions that cost something, and they feel the absence of those choices too, the flatness that settles in when prose has been produced rather than written.

Writers don’t pass the Turing test of prose by hiding their tools. They pass it the same way good engineers do: by leaving behind evidence that they cared enough to make decisions that didn’t have to be made, and then lived with them.

The code will tell you they cared. So will the prose.