The Digital Afterlife

What Happens When Fiction Becomes Training Data?

Writers used to worry about being forgotten. Now, some of us worry about being remembered in the wrong way.

In the past, a story’s afterlife was tied to its readership. A book might go out of print. An essay might fade into a forgotten corner of the internet. But in today’s landscape, a writer’s words don’t just disappear. They’re scraped, stored, indexed, and used—often without consent—to train the very systems that may eventually replace them.

It’s not just about plagiarism or copyright anymore. It’s about presence—about how our most human efforts are being repurposed in the digital ecosystem.

Your Voice, Trained Into the Machine

If you’ve ever published fiction online, there’s a good chance your work is now part of a large language model’s training set. Not because you agreed to it or because you were asked. It’s because that’s how these models are built—by collecting vast amounts of publicly available content to improve their fluency, their tone, their ability to mimic us.

That short story you wrote in 2018?

The LinkedIn post after you got laid-off?

The novel you self-published?

It’s all part of a dataset used to help a machine generate something “human.” Something emotionally resonant, something you would recognize—because it sounds a lot like you.

When Fiction Becomes Infrastructure

This isn’t theoretical anymore. AI-generated content is appearing in query letters, job applications, news stories, and yes—fiction. The emotional arcs we used to craft are now being approximated by code.

Fiction—once considered ephemeral, imaginative, even private—is now a kind of raw material. It's treated less like art and more like fuel. Something to mine.

Consent Was Never Part of the Deal

What makes this moment so complicated is that most writers didn’t sign up for it.There was no terms-of-service box to check. No notice that our work might be used to power the next iteration of generative AI. And even now, most authors have no idea which pieces of their writing have been ingested, or how their words are being used.

Some will argue that this is the cost of publishing online. That anything public is fair game. But “public” doesn’t mean “permission.” And it certainly doesn’t mean “free to be repurposed into someone else’s product.”

This isn’t about gatekeeping creativity. It’s about respecting boundaries. And right now, the industry isn’t concerned about respect. It’s focused on scale.

What We Lose When We Become Data

There’s something sacred about creating a story only you could write. A story that came from your childhood experience, your strange sense of humor, your contradictions and idiosyncrasies. Thia kind of work is the antithesis of automation—it’s the slow, very human process of trying to make order of something messy.

When that work becomes just another data point in a trillion-token dataset, something is lost. Not just ownership. Not just originality. What is lost is meaning.

A Question of Agency

The digital afterlife of our words is no longer a question of legacy. It’s a question of agency.

Do we get to decide what happens to our writing after we share it? Or do we accept that once it’s online, it no longer belongs to us?

These aren’t legal questions, they’re ethical ones. And they’re just beginning to matter in ways most of us never anticipated.

So as AI continues to learn from us, let’s at least be honest about it. Machines may not need permission, but it would be nice if the people feeding them would have the decency to ask.