Jul 28, 2026 • Thinking

Who Wrote This Article?

AI may supply words, structure and speed, but does that make the machine the author?

Table of Contents

by Richard P. Weigand

I have recently encountered a new criticism of writing produced with the assistance of artificial intelligence.

An article is placed into an AI detector, and the detector announces that it was “100 percent AI-written.”

That sounds definitive.

But what does it actually mean?

I write with the help of AI. I bring an observation, an unresolved situation or something that does not fit. I explain what I have seen, why it concerns me and where I think the trail may lead. The AI helps me organize the material. It may question my reasoning, locate supporting evidence, suggest a structure or find clearer words. I reject what misses the point, correct what is wrong and continue directing the work until it says what I intended to say.

The AI does a great deal of work.

But it does not do my work.

Without me, the article would not exist. The AI did not notice the original situation. It did not become concerned about it. It did not decide that the subject was important. It had no purpose to communicate and no audience it wanted to reach.

It did not originate the intention.

That distinction is being lost in the argument over AI-produced writing and art. We are confusing the performance of labor with the source of the work.

The Power Saw Did Not Build the Cabinet

I have spent much of my life around woodworking. There have always been purists who believed that genuine craftsmanship required hand tools. A power saw, planer or router somehow diminished the authenticity of the work.

But a power saw never built a cabinet.

The craftsman conceived it. He decided what it was for, selected the wood, established the dimensions and judged every cut. The machine supplied power, speed and precision. It extended the craftsman’s ability, but it did not supply his intention.

A modern cabinetmaker may perform far less physical labor than a craftsman did two hundred years ago. That does not mean the table saw designed the kitchen.

The same principle applies to a camera. The camera captures far more visual information than the photographer could draw by hand, but the camera does not decide what deserves to be seen. A word processor corrects spelling and moves entire paragraphs without requiring them to be retyped, but we do not therefore credit Microsoft Word with writing the book.

AI is a more advanced tool because it can propose language, structures and even ideas. That makes its contribution greater and the boundary less obvious. It does not eliminate the boundary.

The question remains: Who supplied the purpose?

What AI Actually Does for Me

In my work, AI performs several valuable functions. It helps me organize ideas that arrive in clusters. It shows me where a reader may lose the trail. It can compare sources, test an argument, shorten a passage and offer different ways to say something. It remembers decisions we have made about style and purpose. It is an extraordinarily capable intellectual tool.

Sometimes it produces a sentence better than one I would have written. Sometimes it sees a connection I had not yet expressed. Sometimes it is simply wrong.

I remain the one who must recognize the difference.

That judgment is part of authorship. So are selection, rejection, correction and responsibility. I decide whether the finished article represents what I know and believe. My name goes on it because I stand behind it.

AI cannot do that. It cannot accept praise, suffer disgrace, defend a conclusion or assume responsibility for harm. It cannot care whether the article helps or injures anyone.

It has no personal stake in the truth.

Not Every Use Is the Same

There is an honest distinction to make. Someone can type, “Write me an article about morality,” accept the first result and publish it without examination. That product can reasonably be called AI-generated. The person supplied the request and the decision to publish, but very little of the intellectual substance.

That is not the same process as bringing years of observation, a developed principle, original research and a specific argument, then using AI to help shape the resulting communication.

Calling both products “AI-written” conceals more than it reveals.

The proper question is not whether AI touched the words. The proper questions are:

Who originated the idea?

Who established the purpose?

Who directed the development?

Who exercised judgment?

Who takes responsibility for the result?

Those questions identify the author far more accurately than counting which hands, keys or computer processes assembled the sentences.

The Detector Cannot Know

An AI detector sees only the finished arrangement of words. It was not present for the conversations, corrections and decisions that produced them. It cannot see the original observation. It cannot know who supplied the reasoning, rejected the errors or directed the revisions.

It is examining surface patterns and making a statistical estimate. It is not discovering authorship.

The limitations are substantial enough that OpenAI withdrew its own text classifier because of poor accuracy. Independent studies have also found serious failures under real-world conditions, including false accusations against human writers. A machine that examines patterns in prose cannot reconstruct the chain of intention that caused the prose to exist.

“100 percent AI-written” therefore does not mean what it appears to mean. At best, it means the wording resembles patterns the detector associates with machine-generated text. It cannot mean that AI supplied 100 percent of the thought, purpose, judgment and responsibility.

It has no access to those things.

Where Does Creation Begin?

This argument finally brings us to a much older question.

Where does creation begin?

It does not begin in the hammer, the saw, the paintbrush, the camera or the keyboard. It does not begin in the muscles that perform the labor. It begins with an idea joined to an intention.

Machines do not originate purposes. Left alone, they do nothing. They become active when a living being selects them, directs them and puts them toward an end.

AI has greatly increased what one person can accomplish. It may become the most powerful intellectual tool we have yet constructed. But increasing the power of a tool does not turn the tool into the source.