Distinctiveness and Provenance Are Different Things

Imagine writing an article based on an idea you have been developing for years.

You use AI to help organize it, tighten the argument, perhaps rewrite a few awkward sections. Then you run the finished piece through an AI detector.

The verdict: suspiciously AI-written.

So you start changing sentences.

You make the prose less polished. Swap words. Break patterns. Add little irregularities until the detector becomes happier.

And then I want to ask:

What, exactly, have you improved?

Perhaps you've changed something about the apparent provenance of the prose.

But have you made the article more useful? More original? More conceptually precise? More difficult to substitute?

Not necessarily.

This is where I think much of our conversation about AI-written content has become confused.

We are collapsing two very different questions into one.

Provenance and Distinctiveness Answer Different Questions

The first question is about provenance:

How did this prose come into existence?

Was it written entirely by a human? Generated by AI? Dictated and rewritten? Heavily edited? Some combination?

The second question is about distinctiveness:

How substitutable is the thinking?

Does the piece contain observations, distinctions, knowledge, interpretations, or connections that make it meaningfully different from thousands of other competent pieces on the same subject?

Those questions can overlap in particular circumstances.

But they are not the same question.

Provenance describes the production process. Distinctiveness describes the thinking produced.

Or, more sharply:

Authorship tells us how the prose came into existence. Distinctiveness tells us how easily the thinking could be replaced.

AI-Written Does Not Mean Generic

Once we separate these dimensions, four different possibilities become visible.

Something can be human-written and generic. A person can thoughtfully type every word themselves while expressing ideas that are already familiar, widely available, and easily substituted.

Something can be human-written and distinctive. The prose may be produced entirely by the author, while the thinking reflects unusual observations, original distinctions, deep experience, or a recognizable way of seeing.

Something can also be AI-written and generic. Ask an AI system to produce an article on a familiar topic with little meaningful intellectual input, and the result may be perfectly competent while adding very little that is difficult to find elsewhere.

But there is a fourth possibility our current conversation often struggles to account for: something can be AI-written and distinctive. An expert can bring years of observations, a developed model, unusual distinctions, counterexamples, proprietary language, and a carefully reasoned argument, then use AI extensively to help express that thinking in prose.

The important point is that the production method does not determine the intellectual distinctiveness of the result.

Two pieces can both be human-written and differ enormously in the value of their thinking. Two pieces can both involve substantial AI assistance and differ just as dramatically.

Provenance and distinctiveness are independent dimensions.

AI Is Moving Scarcity Upstream

For a long time, competent prose was relatively expensive.

Writing required time. Structuring an argument required effort. Editing required skill. Those constraints limited how much reasonably polished material could be produced.

AI has dramatically lowered that production constraint.

Competent sentences are becoming abundant.

So are competent summaries, explanations, outlines, and variations on established ideas.

That doesn't make writing irrelevant.

It changes where scarcity lives.

Increasingly, the scarce material sits farther upstream:

What did you notice?

What do you know because you have actually done the work?

Then what distinction did you make?

What question have you stayed with longer than most people?

What contradictions have you observed?

And what connection do you see that isn't obvious?

What are you willing to conclude, and what are you unwilling to oversimplify?

AI can participate heavily in turning those things into sentences.

It cannot retroactively give you twenty years of paying attention.

Your Intellectual Signature Lives Upstream of the Sentences

This is one reason I think intellectual signature becomes more important as AI becomes more capable.

Your intellectual signature isn't your sentence structure.

It is the recurring pattern underneath your work: what you notice, question, distinguish, connect, and interpret.

Over a substantial body of work, those patterns become recognizable.

Perhaps you continually look upstream for the systems producing visible problems.

Perhaps you make unusually precise psychological distinctions.

Or perhaps you connect ideas across fields other people keep separate.

Perhaps you repeatedly challenge assumptions that have become invisible through familiarity.

Those movements of mind can survive changes in format, tone, editing, and production method.

AI might help express them.

It doesn't follow that AI originated them.

That distinction matters.

Provenance Can Still Matter

None of this means provenance is irrelevant.

There are contexts where how something was produced matters enormously.

Journalism may require transparency about reporting and source material. Education may care whether a student's work demonstrates their own learning. Copyright raises questions about source material and ownership. Particular platforms, professions, clients, or institutions may establish disclosure requirements.

Trust can also depend upon knowing how work was created.

Those are legitimate questions.

But they do not turn provenance into a reliable measure of intellectual distinctiveness.

Human-produced does not mean intellectually distinctive.

AI-produced does not mean intellectually generic.

We need enough conceptual room to hold both truths.

The Better Question Is Becoming Upstream Provenance

There may, however, be a more interesting form of provenance emerging.

Instead of asking only:

Who produced these sentences?

We might also ask:

Where did this thinking come from?

Did it originate in firsthand observation?

Years of practice?

Original research?

A distinctive synthesis?

Repeated engagement with a difficult question?

A framework someone developed because existing explanations were insufficient?

A pattern visible across an entire body of work?

This is intellectual provenance.

And I suspect it will become increasingly valuable.

When polished expression is scarce, prose itself carries considerable signal.

When polished expression becomes abundant, we have to look farther upstream for evidence of meaningful intellectual contribution.

Distinctive Thinking Becomes More Valuable When Expression Becomes Abundant

There is an irony in all the anxiety about AI making content generic.

AI may actually increase the value of distinctive thinking.

If everyone can produce polished prose, polish differentiates less.

If everyone can generate a competent explanation of an established idea, competent explanation becomes more substitutable.

The advantage moves toward the intellectual material underneath.

The accumulated observations.

Judgment.

Distinctions.

Questions.

The recognizable intellectual signature.

So I am less interested in whether every sentence can be proven to have passed directly from someone's brain through their fingers and onto a keyboard.

I am interested in whether there is a mind behind the work worth recognizing.

The important provenance question may eventually move upstream.

Not simply:

Who wrote these sentences?

But:

Where did this thinking come from?

That is a much more interesting question.