The Secret Editor Confession: How AI Could Make or Break Your Next Book—and What You Must Know Before Writing a Word

A major publisher recently sent me two nonfiction manuscripts for editing, and both were clearly written by artificial intelligence.

Yikes. Now what?

I don’t want to be the AI police. It’s hard enough wearing my grammar police hat. But I do see a role for AI in writing and in editing. And Medium has weighed in on only publishing human-written posts.

I did what any curious editor would do.

I immersed myself in an AI course for editors so I could appropriately respond. What is my role as editor when I encounter AI-written material? And then what should authors do?

The brilliant Erin Servais, in her course, moved me from skeptical/naïve/uninformed about AI in writing to enthusiastic about using AI in editing and limited use in writing. Here’s how that works — at least right now in book publishing because this whole AI revolution is evolving.

Consider manuscript 1: “No way I used AI.”

The author had been sending me chapters and versions for several months. A nearly final version of a nonfiction first-person type of memoir contained a reflective chapter at the end that read nothing like what came before. It was punctuated properly and contained words the author had not used before. All red flags.

An AI checker (well, two checkers) agreed it was 95% AI written. You can do a Google search (or an AI search) to find AI checkers.

I think we can all agree that AI checkers online are not completely accurate, but their scans can raise the bar for suspicion on AI-written material, and there are plenty of posts here on Medium about words and phrases often used by AI.

An education colleague told me professors are rejecting any essays, papers, and theses even hinting at AI writing. Students, desperate to show they actually wrote a work, are videoing themselves and their computer screens showing them writing.

That’s pretty desperate, but until we have a way to prove something was human written . . .