✪ Who Is the True Author When Writing with AI? - #79

A four-question test that lets any expert writer know whether they're genuinely the author of AI-assisted work.

The Argument

There is a precise, testable standard for when you can consider AI-assisted writing genuinely your work. Even if the AI generated the text, you’re the author. It comes from the work of three philosophers I highly respect.

---

That Strange New Feeling

You’re writing with AI. But there’s a lingering sensation that stays with you every time you let AI drive the majority of your interaction with it.

It's more like you're not sure what you just did. You contributed something, clearly. But you’re not sure if that something is enough to call yourself the author. And nobody around seems to be addressing this issue out loud, if not to discard it as irrelevant or to condemn it as if it were an unforgivable war crime..

1) Distant Writing

What do you do when you write with AI?

You don’t just type a prompt and paste the output. If you are an expert, you think about the problem first. You shape and give direction to the question. You push back and counter many times on what the AI suggests and recommends. You cut, re-order, drop, reframe, question and verify, and add things only you would know. You assemble something that feels, in the end, like it came from you, even if the actual sentences didn’t.

Luciano Floridi, an Italian philosopher who studies how technology changes what we do and who we are, looked at this exact process and gave it a name: distant writing. You’re writing, but at a distance from the sentences themselves. You’re the one making all the meaningful decisions but you're no longer the one typing every single word.

Like a museum curator who doesn’t paint the paintings though the exhibition is entirely hers. She chose what to include, decided the sequence, wrote the frames, made it mean something. The visitor leaves having had her experience of art, not just of having seen a random set of paintings.

But what Floridi uniquely observed is what distinguishes 'distant writing' from simply prompting and delegating to AI the writing of anything. Distant writing is clearly recognizable from delegating to AI (failure mode) because it is characterized by an author carrying out four specific curatorial decisions:

.

a) Framing the Problem (architectural)

You thought carefully about what question this text was going to answer, and it was a question only you were positioned to ask, based on your knowledge of your readers and your field.

Failure mode: You gave the AI a topic. The AI framed the question. You accepted that framing. The text answers something nobody asked you specifically to answer. It’s a good article about your topic, but it isn’t quite yours.

.

b) Steering the Exchange (dialogical)

The conversation with the AI went somewhere specific because of what you know. You pushed back on things that weren’t true to your experience. You redirected toward what your specific readers actually need. The AI’s output changed because of your input in ways it couldn’t have changed without you.

Failure mode: You ran the exchange on the AI’s logic. You accepted what it thought was relevant. The draft came back and all you did was polish it and make it sound more like you. The article ended up being about what the AI thought was important, not what you thought it was.

.

c) Verifying the Content (evaluative)

You checked what the text says against your own knowledge and stood behind it. Not just: Are the facts correct? But: Is this what I actually believe? Is this consistent with what I’ve learned? Can I defend every claim here from my own experience and understanding?

Failure mode: You fact-checked the surface but not the substance. The text is accurate. But it doesn't reflect your true beliefs, it reflects a plausible version of what a knowledgeable person in your field might say. There’s quite a difference.

.

d) Assembling the Whole (integrative)

You made this cohere with what you’ve been building over time. It connects to your body of work, your developing argument about your field, the specific perspective your readers follow you for. Someone who reads this and your last ten pieces can see the same mind at work.

Failure mode: The piece stands alone. Well-written, well-structured, informative. And 100% promptable. It could have come from any competent writer with access to the same AI and the same topic. Nothing in it requires you specifically to exist.

When you don’t do these four things, the text is not yours.

You are not the author.

You are not the author because the quality of your article isn’t traceable to your specific competence.

Yes, you wrote a prompt, checked and refined, and were present while it was made. But that’s not the same as being its true author.

2) The 3 Archers and the Apt Test

Floridi’s frame is useful, but it leaves the real question open. So you did four curatorial things with an AI. Does that make the text yours? Or did the AI do the actual intellectual work, and you just supervised?

Tiegue Vieira Rodrigues, a Brazilian philosopher, read Floridi and asked exactly that. And to answer it, he reached for a framework, the APT test, built by another philosopher, Ernest Sosa, that turns out to be exactly the right tool.

Sosa’s framework boils down to this:

Imagine three archers.

The first archer shoots an arrow. A gust of wind catches it mid-flight and blows it into the bullseye. Lucky. The shot was accurate, it did hit the target, but it wasn't really his achievement. The wind did it.

The second archer is trained but was distracted when he shot. The arrow hit anyway, through some combination of habit and luck. Accurate, but not clearly apt — you can't trace the success cleanly to the archer's competence.

The third archer is trained. He has technique, muscle memory, years of practice. He shoots and the arrow hits the bullseye because he’s good. That shot is apt. It succeeded because of his skill. It’s his.

Sosa’s point: A performance is truly yours only when it succeeds because of your competence, because you're good at it and the goodness flows from that.

Now apply this to writing with AI.

A text is your genuine achievement and the work is actually yours, when its quality is traceable to your competence, not the AI’s output. When the text is good because of the choices you made, the knowledge you brought, the judgment you exercised.

This is what Rodrigues calls the Apt Curation Model. And it answers the question Floridi left open: yes, the work can be genuinely yours. But only when you can honestly say: this is good because of me, and because of the specific actions I took.

You can actually check.

“…distant writing reframes rather than diminishes the author’s role, shifting epistemic competence from direct composition to second-order curatorial acts:
> architectural design,
> dialogical refinement,
> evaluative verification, and
> integrative synthesis.

Applying Ernest Sosa’s virtue epistemology, we develop the Apt Curation Model, which positions the human author as the central epistemic agent, not as sentence originator, but as virtuous curator whose competence makes the text an apt epistemic performance.

Source: Rodrigues, T. V. “The Apt Curation Model: An Epistemic Virtue Theory of AI-Assisted Authorship” (2026)

3) Authorship - What It Really Is

Authorship has always been about selecting, directing, making hard choices, discarding things. It’s not just about sentence production. Look at how a book is made. Many other people are involved in checking, verifying, and improving your work, but you are the one who is the actual author.

Well, the news is that AI just made that impossible to ignore.

The anxiety most people feel about writing with AI assumes that authorship lives in the sentences, that being the author means being the one who typed the words. But if you look around and think about it, that was never really true.

Ghostwriters have always existed. Editors have always shaped texts profoundly. Researchers have always contributed knowledge that ended up in other people’s books. What we called “authorship” was always really about: whose choices made this what it is?

Authorship was always that, and AI hasn't changed it. It just removed the never-questioned belief that sentence production and authorship were the same thing.

Rodrigues’s standard isn’t a new rule invented for the AI age. It’s what authorship always was, stated precisely.

The author was always the one whose competence made the text what it is. If that’s you, calling yourself the author is 100% honest.

4) Are You the True Author? The Four-Question Audit

Here’s a practical test. Run it on anything you’ve published. Not to score yourself but to understand better what authorship really is.

1. Did I design the question this text was answering, or did I accept a framing that came from somewhere else?

2. Did the exchange between me and the AI go somewhere specific because of knowledge or judgment that only I have?

3. Does this text reflect what I actually believe and stand behind — not just what's accurate?

4. Does this connect to the work I’ve been building, or could it have come from anyone?

Remember: to be the true author of anything you co-produce with AI, you need to have taken actual responsibility for these four actions.

If all four answers are yes, you are indeed the author, no matter how many words the AI has generated and you have left untouched.

If not, now you know where to look to regain your authorship.

Why This Matters for Curators

Three key implications:

1. Curation becomes the core skill of authorship itself.
You’re no longer competing on writing fluency (which AI commodifies) but on taste, judgment, and editorial decisions, the same skills curators have been developing for years.

2. Responsibility, not ownership, is what matters.
Just as Bernini designed St. Peter’s Colonnade without building it himself, the distant writer designs the narrative architecture and remains fully accountable for the final product.

3. Your viewpoint becomes irreplaceable.
The specific lens through which you select, refine, and curate LLM output cannot be faked or replicated by competitors, it’s exactly what makes a curator valuable.

---

Go Deeper

• Rodrigues, T.V. (2026). “The Apt Curation Model: An Epistemic Virtue Theory of AI-Assisted Authorship.” Philosophy & Technology.

• Floridi, L. (2025). “Distant Writing.” Philosophy & Technology.

• Floridi, L. (2008). “Levels of Abstraction”. Minds and Machines.

• Sosa, E. (2007). “A Virtue Epistemology: Apt Belief and Reflective Knowledge” Oxford University Press.

• Sosa, E. (2015). “Judgment and Agency”. Oxford University Press.

---

---

Paradata

Intent / Why: I wrote this article with the goal of clarifying the issue of authorship in the age of AI-assisted writing. Does someone who writes with AI remain the author of what he writes? The answer that emerges is: It depends.

The more apt the content produced and traceable back to the human author’s knowledge, choices and decisions the more clearly authorship remains with him and not with the tools or other human beings through which he has produced what he had in mind. Just like a film director. He is not the actor, not the cameraman, the lighting director or the one who wrote the dialogues, yet at the end of it all, he is the credited author of that opera.

Process: A few months ago, I had saved aside (inside my Sublime digital garden) as possible topics for TRUSTable two academic papers I had found to be truly fascinating for my focus and interests. In those months I had gone back a few times to them to re-read them and better understand their implications, which seemed to me highly relevant for my specific readership and for one of their growing fears: authorship.

And so, a few days back, last Saturday, when I started working on this piece, I brought to the table the academic papers by Tiegue Rodrigues and Luciano Floridi and asked AI to assist me in developing an outline. The goal was to clarify for experts and consultants writing online, the reason and the way out of their frustration about using AI for actual writing.

So I engaged with Claude in making sure that the feeling I myself had when I started to test AI for writing, would be highlighted and be the identification point for the reader. From the rough outline the AI generated, I restructured the order of the sections, eliminated what I considered superfluous (redundant stuff about the polished, shallow writing AI generates - those that look good, but really have no sensible depth or logic). At this point I used once again the outline as a canvas and started writing on it from beginning to end. I kept a few passages generated by AI intact, and I painted over the rest in my own style.

I then let it rest for a night. Picked up the next day and went through it again a second time, doing more edits, dropping anything unnecessary and making sure that my voice was actually in there. I then went back and re-read and verified my starting sources to see if there was any particular point that I could have added or that was missing from this article.

*By doing this I discovered the work of Oliver Ding, which I had forgotten about and which has a very interesting theoretical take on indie creators who systematize and publish their reflective reasoning (why they chose, what they rejected, what the pattern reveals) - essentially, what I am doing here with this Paradata section.

Once again I let the article rest for a full night and then repeated the process on the next day. What you’re reading now is the result of these steps.

Time: Research and study: 2 hrs. Identify specific issue to address, possible title / subtitle and brief AI on it: 1.5 hrs. First review, edit and optimization of rought draft generated by AI: 1.5 hrs. Second manual review: 2 hrs. Third review: 2 hrs. Cover design: 1 hour. Paradata: 1.5 hrs. Formatting: 1 hour. Links and references: 1 hr. Grammar and check verification and corrections: 1.5 hrs. Total time: apx. 15 hrs.

Tools: Sublime, ChatGPT 5.5, Claude Sonnet.

Cover image: Design by Robin Good executed by ChatGPT Imagen 2

---

What I Can Do For You

I help experts and consultants who are building a new online career gain the authority, credibility, and visibility they initially lack.

I do this by teaching them how to become top trusted curators in their area of expertise.

For this purpose I’ve created a free, focused 55-minute video workshop that walks you through:

• Why curated content formats are so useful for building credibility and authority

• When to use them and what requirements they have

• An updated list of real-world examples of curated formats at work

• The specific tools you need to curate

• The actual key steps that transform researching and writing into curation

• The 11 typical mistakes novice curators make

One-to-One Audit & Strategic Advice

For experts and consultants looking for ways to:

a) Create Value and Build Trust for Non-Writers (Curation)
I help you identify and master your own personal ways to gain authority and value by curating insights, research, news, resources and tools in your field of interest.

b) Improve Credibility and Trust
I review your content, positioning and goals to identify the best editorial strategies to build trust and credibility around your focus.

c) Positioning and Personal Branding
I analyze your market positioning to identify key strengths and weaknesses. I help you redefine a strategy to differentiate yourself from the competition while increasing the practical value you bring to your readers.

Available in two tiers: a) for those just starting and b) for those who have been publishing for more than a year, but are not seeing results.

Follow a path with a heart.
The time is now.

From Koh Samui (TH)
Robin Good