The Em Dash Incident

When I returned to long-form fiction after so many years away, I did so with more excitement than hesitation. I’ve lived over half a century now, gathered enough experience to feel like I finally had something worth saying, and when I sat down to write, the pages came quickly. My niche felt clear. My stories followed.

From the beginning, I’ve always been my own first editor. I write freely, then go back with a critical eye—trimming, reshaping, asking myself a simple question: If I were the reader, what version of this would I respect the most?

Somewhere along the way, I added a new step to that process.

I started using ChatGPT.

At first, it was nothing more than a proofreader. I’d feed it pages and ask for a report—grammar, clarity, consistency. I’d already done my own pass, of course, but it offered a second set of eyes that didn’t get tired and didn’t miss much. More importantly, it could be guided. It could be told that voice mattered more than rigid adherence to rules. Sometimes a sentence should end on a preposition. Sometimes rhythm matters more than correctness.

Used that way, it became a tool. A useful one.

And for a while, the process worked smoothly.

Then I became aware of a different conversation happening around all of this.

There’s a growing tendency online to assume that if something is polished, or successful, or even just present, then it must have been generated by AI. The accusation gets applied broadly—good work, bad work, mediocre work. The specifics don’t seem to matter nearly as much as the impulse to label it.

I’d be lying if I said that didn’t get into my head for a bit.

At one point, I went back through the early chapters of my manuscript and removed every em dash I had written. Every single one. Not because they didn’t belong there, but because I had convinced myself someone, somewhere, might point at them and decide they meant something they didn’t.

That didn’t last.

The more I sat with it—especially while working on Beyond That Starry Sky, a story that, ironically, explores AI from a different angle—the more I realized something simple: I know how I write. I know the work I put in. And I wasn’t going to start sanding down my voice to preempt someone else’s assumptions.

So I put them back. Every em dash I wanted.

Because at the end of the day, tools don’t write books. Writers do.

For me, AI remains exactly what it started as: a tool in the editing process. It doesn’t generate my stories. It doesn’t decide my characters. It doesn’t shape my themes. What it can do is offer perspective—sometimes useful, sometimes not. Like any tool, it requires judgment. Some suggestions get ignored. Others help me see a sentence more clearly and refine it.

That’s the extent of it.

There’s also a broader argument that comes up often—about AI being “trained” on the work of others. And I understand why that raises concerns. But I also think it’s worth remembering that every writer, in a very human way, is shaped by what they’ve read.

My own voice didn’t emerge in a vacuum. It was formed over decades of reading authors like Richard Bach, Dean Koontz, David Gerrold, Dan Brown, Larry Niven, J. R. R. Tolkien, and George R. R. Martin, among many others. Every writer carries those influences, whether consciously or not. That’s part of the craft. It always has been.

None of this means every use of AI is good. It isn’t. If someone chooses to flood the market with low-effort, generated work, readers will notice. And readers will decide what’s worth their time.

They always do.

But for those of us who are still putting in the hours—late nights, early mornings, clacking away at keyboards, shaping sentences one at a time—the presence of a tool in the background doesn’t change the heart of the work.

The story still has to be written.

And I know exactly who’s writing mine.

Until next time.