I mentioned in my last post that, during December 2025, I made a change to my writing process that has paid off in spades. It came after a lot of serious thought and consideration, but now that I’ve made the plunge, I can’t believe how much difference working with an AI has made.
There’s a lot of yelling and screaming out there, these days, about whether AI writing is “real writing,” and I should make one aspect of my position on the matter clear: if the writing itself is generated by AI, it’s not the author’s work. This is evidenced by the fact that, if text is solely written by generative AI, it cannot be copyrighted.
Further, I still fully agree that the use of copyrighted work to “train” a LLM is unethical if the author hasn’t given permission — and in my personal case, I have explicitly prohibited the use of my work for such purposes. This holds true even for my fan fiction, as certain aspects of fan fiction are, in fact, copyrightable to the author and not the source material’s creator (and thus, it is possible to plagiarize fan fiction).
So why did I “cave” and begin using AI?
Because my roadblocks weren’t in the actual writing.
My difficulties had to do with both pre-writing and editing. Sometimes my plots were internally inconsistent or depended a bit too much on coincidence. Sometimes my characters were flat to the point of becoming caricatures. And sometimes my dialogue sounded all the same, even when multiple characters were speaking. Even worse, I’m well aware I have a tendency to overuse words and phrases, and my writing isn’t always as “punchy” as it can be.
It began when I copied-and-pasted a short scene (about 500 words) into a GPT application and asked it to evaluate, not for the event being described, but for the quality of the writing. Even without me having “trained” the LLM, it gave me a clear and useful critique. I was able to use that to improve the passage considerably and it showed when I later took it to a live writing workshop.
Read that paragraph again: I didn’t ask the AI to rewrite the passage for me. I asked it to give me feedback on something I had already written and then I used that feedback to improve it myself.
That’s not using AI to write. That’s directly comparable to using it as an editor and/or a beta reader, depending on what you’re writing.
Switching to AI-assisted writing is proving to be a gradual process for me, with a noticeable learning curve. As of right now, I’m still on that curve, although I’ve already learned how to organize my chats and information to make them easier to reference in the future. I have also — and this is big — actively turned off any aspect of this particular application that would allow the use of my work to train the LLM for others’ use.
That’s no guarantee it won’t happen, of course, but one of the benefits of being a lesser-known author is, quite honestly, nobody cares enough to want to imitate my style or my work. For now, I’m continuing to use AI, but I am doing so in a manner that’s cautious, thoughtful, and always with the same end goal in mind: improving my own writing, not outsourcing it to a machine.