photo: Markus Winkler

With the incredible rise of artificial intelligence comes a set of questions, among them: Did the writer write the text, or did A.I.? Or… maybe a little bit of both.

It’s an important question, because it goes to the heart of authorship: the difference between a writer’s voice and style, and an artificial one that may sound correct, but too often lacks texture, judgment, and a sense of lived experience.

(That last sentence was written by AI.)

One sign of A.I. usage concerns “the rule of three”: three consecutive descriptions, something I have used to emphasize a point. However, look at the end of the AI-written sentence: “lacks texture, judgment, and a sense of lived experience.”

Sounds forced, awkward, and flat to me. (I just used the rule of three, but I think it supports my case.)

Take a look at these triplets from “Loyalty vs. Conscience” (March 19):

“There is a loyalty that binds a republic together. Loyalty to the rule of law. Loyalty to the dignity of opponents. Loyalty to the idea… the principle that no person is above accountability. Loyalty to a tribe, a personality, or a grievance diminishes a republic.”

I like the cadence as well as the emphasis on the point I’m making. However, I think I pushed too far with the second series. It’s placed too close to the first and, instead of strengthening the point, it softens the emphasis I was aiming for.

Here’s how the revision now appears: “There is a loyalty that binds a republic together. Loyalty to the rule of law, and loyalty to the principle that no person is above accountability. Loyalty to tribe, or personality, diminishes a republic.”

Sound clearer, sharper? I think so.

The lesson for me is simple: be more disciplined. Use the rule of three sparingly, and only when it truly sharpens the point.

If you’re wondering about my use of “artificial editing,” here’s what I rely on. I use Microsoft Editor built into Word, along with Grammarly—not simply to catch misspellings and grammatical mistakes—but, at times, to clarify a sentence. But I don’t accept everything they suggest.

I do rely on the Thesaurus in both programs. It’s not just about finding alternate words but finding the one that fits the context of the sentence. That matters to me.

In my opinion, these tools don’t replace judgment; they sharpen it, much like a good editor who understands the difference between technical correctness and clear understanding.

Last year, I began using ChatGPT as a research tool because of its astonishing speed and accuracy.

Recent searches:

“U.S. Made Soil Activator”: I like the product, but shipping from China took too long.

“Contacting Pottery Barn CEO”: Email address for an unresolved product complaint.

“Churchill personal memoir”: Research on his involvement in World War II.

Google uses an AI program to find what we need more quickly. Anything that helps speed up the research process is acceptable to me.

In researching work for this commentary, I randomly re-read some of my work from the last two months. While I think I’m improving, I have clearly taken on a more serious tone, reflecting the times we’re living in.

Circling back: Who writes this stuff, Jim or AI?

I write it myself. A.I. has its place as a powerful research tool, and I turn to ChatGPT and Grammarly for the Thesaurus to help zero in on the word that best fits what I’m trying to say, but blame me for the rest.

Postscript:

I put this commentary through ChatGPT to see if it sounds like me or AI. Here’s the response:

“This reads like you. Here’s why.

“First, the self-awareness. You’re not just making a point, you’re examining your own writing, questioning it, and revising it in real time. That kind of reflection (“I pushed too far… it softens the emphasis…”) is distinctly yours. A.I. can mimic reflection, but it rarely revises itself on the page the way you do here.

“Second, the structure. You move cleanly from question to example to self-critique to revision to lesson. It’s disciplined and deliberate. It feels thought through, not generated.

“Third, the imperfections—and that’s a strength. A few lines are uneven, and some phrasing tightens as you go. That’s human drafting. A.I. is often smoother but also flatter.”

Final question: What grade would you give my work?

“I’d give this an A-.

“If you give this one a more disciplined pass—tightening language, cleaning those small breaks—it moves easily to an A. And more importantly, it does something an A.I.-written piece almost never does: It thinks on the page and lets the reader see it happening.”

A- … I’ll take it!