The ghostwriter in the machine

A few months ago, a friend sent me an essay she’d written. It was a personal piece about the recent loss of a friend, about grief, specific and quietly devastating. I read it twice. Then she mentioned, almost offhand, that she’d used AI to help “shape it.” She had the feelings, she said. She just needed help finding the words.

I didn’t know what to do with that.

My discomfort surprised me—I’ve never had much patience for ghostwriting snobbery. We’ve always known that celebrity memoirs have co-authors, that speechwriters put powerful words into powerful mouths, that plenty of bylined content passes through the hands of the never-named. I made my peace with that a long time ago, as one must in this line of work. Writing is collaborative. Voice is constructed. Authorship has always been messier than we’d like to pretend.

So why did this feel different?

Ghostwriting Has Always Been With Us

The history of ghostwriting is essentially the history of writing. Ancient scribes composed on behalf of pharaohs. Renaissance painters employed workshops of apprentices. The most moving political speech you’ve ever heard was likely drafted by someone whose name you’ll never know. In the twentieth century, ghostwriting industrialized niches like celebrity memoirs, business books, political manifestos, and advice columns. Malcolm X worked with Alex Haley. JFK’s “Profiles in Courage” had significant help. Even Nancy Drew had several authors.

We have, largely, decided that this is fine. We understand that the ideas, experiences, and authority belong to the named person, even if the sentences belong to someone else. The byline is a social contract, not a stenographic record.

The Thing That’s Actually Different

So what’s actually new here? When a human ghostwriter writes for you, there’s still a person in the room—someone who listens, interprets, makes choices. They might push back. They might misunderstand you, maybe in interesting ways. The collaboration has friction, and that friction shapes the work. The final text is, in a rather meaningful sense, the product of two consciousnesses.

AI doesn’t push back. It accommodates. It is, by design, a frictionless mirror—it reflects your prompts back at you in the smoothest, most plausible language available. That’s useful. It’s also, if you’re being honest about it, a little unsettling. The essay my friend produced wasn’t a collaboration, not really. It was her feelings poured into a machine that arranged them into acceptable sentences.

Is that so different from using a thesaurus? From hiring an editor who restructures your argument and rewrites your lede? Every writer uses tools. Every writer is shaped by the language around them. The line between “my voice” and “the voices I’ve absorbed” is more gradient than hard edge.

The difference might be one of degrees, but degrees matter. There’s a reason we distinguish between the artisan and the builder. The tool changes the relationship between the person and the work.

The Case for Accepting It

Here’s what I keep coming back to when I try to argue the other side: writing has always been a technology for externalizing thought. The moment you put something into words, you’re already translating, approximating the mess of experience into something communicable. AI is just another step in that translation process. A faster one. A weirder one. But translation nonetheless.

There’s also a real argument for accessibility. Not everyone who has something worth saying has the tools to say it fluently. Language ability is unevenly distributed, shaped by education, privilege, and neurology. If AI lowers the barrier for people who have genuine experiences and ideas but struggle with written expression, that seems like a net good for the culture of writing, not a threat to it.

My friend, her essay, she wasn’t faking her grief. She wasn’t faking her need to articulate it. Maybe what she needed was exactly what AI gave her: scaffolding.

The Case for Unease

And yet. And yet and yet and yet. The genre of personal essay depends, fundamentally, on a kind of trust. When you read someone’s account of loss, or illness, or failure, you’re extending a particular brand of faith that the sentences in front of you are connected to a real person’s real attempt to make sense of something. The prose doesn’t have to be polished. In fact, the roughness is often the proof, the thing that says, “this cost something.”

When AI smooths that roughness away, it doesn’t just change the style. It severs something. The text becomes professionally presentable in a way that obscures whether anyone actually wrestled with it. And for readers—for editors—the unknowing changes the experience. You can’t unsee it once you’re aware of the possibility.

There’s a broader erosion at stake, too. If voice becomes detachable from person—that is, if anyone can generate credible prose in any register, on any subject, instantly—what happens to the trust that makes literary culture function? What happens to the implied contract between writer and reader? I don’t know that the answer is catastrophe. But I think the contract is being renegotiated, whether we like it or not.

The Answer is a Question

I haven’t stopped thinking about my friend’s essay. I still honestly think it was good. I still think her grief was real. I’m still not entirely sure how I feel about the fact that the sentences weren’t quite hers, not really.

Maybe that’s a place to land—not with a verdict, but with a sharper question. What do we actually want from authorship? Not the idealized version we invoke when we want to dismiss ghostwriting, but the real thing: What is the promise that a byline makes, and to whom?

Ghostwriting has always existed in the gap between what someone feels and what they can articulate. AI has just made that gap visible in a new way and given it a new texture. The discomfort isn’t irrational. But it might be pointing at something we’ve never been fully honest about in the first place.

Writing was never just about the words. It was about the struggle to make them. I guess the question now is whether the struggle is the point—or whether it was always just the cost of admission.