The Biggest Lie About AI Writing

Over the past year I've had more conversations about artificial intelligence than I can count, and the striking thing has never been the disagreement. People who agree on nothing else keep arriving at the same worry: what is AI doing to writing? It's a fair question. It's also the second question ;and it took me longer than I care to admit to notice that no one, myself included, had stopped to ask the first one. What is writing actually for?

For centuries we answered that without noticing we were answering. We measured writing in keystrokes and visible effort, credited authorship to whoever typed the words, and slid into the assumption that the point of writing is to produce text, which is like assuming the point of running is to wear out shoes. Nearly every argument about AI writing rests on that assumption whether it acknowledges it or not. Writing was never about producing text. Long before it produces a page, it produces the writer.

This is not a comforting thing to say in a room full of people worried about their jobs, so let me be plain about the stakes. If writing exists to produce words, then artificial intelligence changes the economics of writing forever. Machines already generate paragraphs in seconds and get faster every month; if the job is to fill the page, we have lost it, and we should all be updating our résumés. But that was never the job. Every difficult sentence is a decision. Every revision forces the question of what you actually mean instead of what you'd like to sound like. Every argument drags your unexamined assumptions into the light and asks you to defend convictions you can't yet defend. The struggle is never with the keyboard. It is with yourself, which is exactly why we'd love to blame the keyboard. Writing matters because it leaves us changed. The paper is evidence. The person is the outcome.

Most of that happens where no one can see it, the false starts, the abandoned paragraphs, the twenty minutes spent staring at a wall for one honest word. That is not the overhead of writing. That is the writing. It's why two people can turn in essays of identical polish having lived completely different experiences: one arranged facts he already had, the other clawed toward an idea that rearranged him on the way. The pages match. The people don't. We would never celebrate an app that let a pianist skip the ten years and be brilliant on demand, because the ten years were not in the way of the musician. The ten years were the musician. Medicine builds the physician, teaching builds the teacher, decisions build the leader. Every practice worth having manufactures the person performing it. We assumed our work was producing our output. It was producing us.

Here is the objection that deserves respect, because it's a good one. AI does not feel like spell check or a patient editor tidying up work you've already done. It hands you a finished draft, a whole argument, a passable imitation of your own voice, before you've had a single real thought about the subject. That is what makes this moment different, and the unease is not paranoia. The danger was never that the machine writes faster. The danger is that it offers to skip the part that was doing all the work on us. Take that seriously. AI reaches deeper into the act of creation than anything we've used, and it produces language with a fluency that should unsettle you. But none of that answers the question; it sharpens it. The question was never whether a machine can produce words that resemble ours. It's whether we'll keep doing the human work that gives those words a reason to exist. Technology decides how we compose and send. It does not decide what is true, what is worth saying, or who we become while saying it. That was never on the table, and it still isn't.

None of this makes AI merely the next word processor. The word processor changed how fast the words arrived; this changes how easily the thinking can be skipped and treating them as the same underestimates both. But raw capability was never what made someone a writer. Authorship is a matter of who chooses, who judges, and who answers for the result and by that measure, however fluent it becomes, artificial intelligence is the next writing instrument, not the next writer.

And yet notice how selectively we moralize about all of this. We have made our peace with a staggering amount of help. Editors rewrite our clumsy sentences and earn a warm line in the acknowledgments. Mentors hand us the idea and let us believe we found it ourselves. Spell check quietly saves us from looking illiterate, and a search engine answers in seconds what used to cost an afternoon in the stacks and none of it troubles our conscience for a moment. Then one more form of assistance arrives, and suddenly it is a moral emergency. We shame the student who admits to it, side-eye the colleague who uses it, and demand disclosures we would never dream of requiring for a good editor. If the real concern were skipping the formative work, we would ask whether the work got done. Instead we condemn the tool for existing. That should tell us the panic was never really about ethics.

Because it was never a fight about software. Writing became an identity. We spent years learning to think on the page, collecting the degrees and the credibility and the quiet certainty that we were, whatever else, people who could write and an identity built on what you can do turns fragile the instant something else can do part of it. This is what every serious technology does: it doesn't just change what we do, it exposes what we believed. The printing press didn't only mass-produce books, it dismantled who got to be an authority. The telescope didn't only find stars, it demoted the Earth. Photography broke painters, the calculator convinced a generation arithmetic was finished, and every time, the tool forced a nervous species to ask where its worth had been hiding. AI is our turn. Underneath the arguments about ethics and plagiarism policies sit the questions no one says out loud: if a machine can do the thing that made me feel exceptional, what makes me exceptional? If intelligence is cheap, what's still valuable? If the work that earned me admiration now takes nine seconds, where did my worth go?

Those aren't technical questions, which is why the debate keeps costuming itself as one. It's easier to argue about disclosure rules than to admit you're afraid, and easier to call a tool immoral than to ask why it got under your skin so fast. The loudest reactions say less about the technology than about the self-image we're defending. That's not a failing; it's a reflex as old as tools. Every generation has had to relearn that its dignity was never stored in the tasks it happened to be good at.

So here is the lie, the real one, under all the others. Not that machines can write, they can. Not that using them is cheating, or that authorship is now impossible to define. The lie is the belief that writing exists to produce text. It never did. The page was always just the receipt. The transaction was happening inside the person holding the pen. The first time I wrote that sentence I stopped typing for a while, because it named something I'd felt for months and hadn't been able to say.

And it doesn't stop at writing. Every real practice leaves two things behind: the thing anyone can see, and the person who got rebuilt making it. The first is easy to measure and, now, easy to counterfeit. The second can't be downloaded, delegated, or generated on demand, and it's the only one that lasts. Teaching, leading, healing, creating, the ordinary work of an ordinary day, each produces something visible, and each is quietly producing someone you can't see. That is what AI has made impossible to ignore, and it will matter more in the months ahead than any argument about prompts or citations.

It also dissolves the shaming, because the question was never which tools touched your work. It was whether you did the work of becoming someone, and no disclosure statement ever written can measure that. Which is where someone always objects: then how can we tell? If we can't police the tool, how do we know who actually did the thinking? That only sounds hard because we've been answering it backward, auditing the text for the machine's fingerprints, running it through detectors wrong often enough to end careers, demanding confessions about process. We are interrogating the algorithm when we should be evaluating the person. A borrowed page is easy to produce and, given a year, impossible to catch. A formed mind is neither. Ask someone the second question, the one past the summary. Ask them to defend the idea, to carry it somewhere new, to think in front of you. Insight that was never earned buckles at the first real push, and judgment cannot be faked in real time. The thing worth measuring was never in the document. It was always in the person, their depth, their discernment, the life behind the words and that is what we should have been evaluating all along.

There's an irony worth sitting with here. AI may make authentic human authorship easier to recognize rather than harder. When everyone can summon polished language on demand, polished language stops being impressive, and what begins to matter is the quality of the mind behind it; insight over fluency, judgment over grammar, wisdom over word choice. The shortcut can imitate the appearance of writing. It cannot substitute for the slow formation of a person who has something worth saying.

The age of artificial intelligence is going to ask a great deal of us, starting with the oldest question we have: what kind of people do we intend to become while using what we build? Before we decide what the machines should do, we'd better remember what writing was always doing to us. That's where this begins, not with the technology, but with the person holding the pen. It's where my own thinking started. I suspect it's where yours will too.

Watch On YouTube https://youtu.be/nul3nhxJi90?si=pU3WOBNvqMYXZUxS

Listen of Podcast: https://open.spotify.com/episode/13FI5lghl4qbvw4u4ZkxSs?si=cG92gxGITGeUwK_A4V8a3w