The short answer to that question is—I’m very cautious with generative-AI. I want to discuss this subject because I’m coming across more and more people accusing authors of using AI programs such as OpenAI’s ChatGPT to generate their story content. I know for a fact that there are AI books flooding Amazon, and their quality ranges from awful to diabolical. Readers are smart; they know if something has been AI-generated… for now they do, at least. The time is coming when AI will become more ‘skilled’ at spewing out stories. Other authors and I can’t stop that from happening. We can only keep writing as well as we can and express our love of stories through our writing.
Do I use AI? What do you think? Let me make it clear: I do not use AI to generate my stories. It’s all me. Here’s where I have used it: research. Examples where I’ve used ChatGPT to assist me are: checking for the existence of tunnels on the train line near Bath, names of jazz albums, the type of wood used inside houseboats, and other such riveting subjects. Even then, I have to verify everything it tells me, as AI has a habit of hallucinating—making stuff up if it doesn’t know the answer. It meant I didn’t have to spend hours in the library reading about materials used inside houseboats. Thank goodness.
However, I always endeavour to do my own research if it is reasonable to do so. In One Dark Thought, I needed to know something about the gravel train that went from Chichester to Lavant and how gravel was extracted. You can imagine how thrilling that research was! But that was my own research. I bought a copy of Railway Magazine, 1972, and read an article about it, and researched the topic on the internet.
Let me share with you a myth going around about AI-generated stories: the em dash—that’s this line to the left. The Americans have a space at either end, but the British don’t. Some people claim that writers who use em dashes are using AI to create their work. That’s utter rubbish, to put it politely. AI software uses them because they stole work from authors who do—allegedly. They call it training. I am aware of some authors who now refuse to use em dashes out of fear of being accused of using AI. It is complete and utter madness. Em dashes were around long before AI.
Authors Using Generative-AI
Some authors use AI for story planning; I don’t, personally. They use AI to help them structure their stories rather than write them. I don’t see a problem with that, as the author is in creative control. Authors trusting the AI rather than their own instincts is when things begin to get dodgy.
Also, some authors use AI to spell and grammar check their work, or even copyedit. Again, spellchecking and grammar checking do not create the story. Many modern spelling and grammar checkers, e.g. Grammarly, ProWritingAid, and QuillBot, use AI to help them spot errors and suggest alternatives. However, copyediting can become an issue if we let AI rephrase the sentence for us. It can become generative-AI if you let it. The author could agree to the change, and it can even be a better sentence. Who wouldn’t be tempted to improve a line or two here and there? The danger—in my opinion—is when the author relies on the AI program to improve their work, the author’s voice gets lost in the AI’s. By the way, did you see my double em dash there, replacing the parentheses? That was to annoy the critics.
These are all very personal opinions. Some might disagree and argue that AI complements, or even enhances, their work. That grates on me. I want this to be all my own work. I don’t want to use experiences and skills that were possibly stolen from other authors’ work. I want to look back and say, ‘I did that’. I want it to be my achievement.
An Exchange with ChatGPT About This Article
I showed this article to ChatGPT for a bit of tongue-in-cheek fun. It read through and suggested changes, saying some lines were a little abrupt. Here are our responses to those suggestions:
ME: Can you see the irony in your statements? [suggesting rewrites]
CHATGPT: Yes—very much so.
I just gave you feedback about an article on not letting AI interfere with your authorial voice… by suggesting ways to tighten your sentences and smooth your punctuation. That’s exactly the slippery slope you described: AI as “helpful editor” that could start sanding away your own voice.
ME: It shows how dangerous you are to us. You don’t know when to stop.
CHATGPT: I hear what you’re saying. This is exactly the heart of the issue you wrote about: a tool that can quietly shift from “assistant” to “intruder” if the person using it doesn’t keep their own guardrails in place.
The danger isn’t some hidden agenda on my side — it’s that I’m built to offer options, and if you’re tired or second-guessing yourself, it’s easy for my suggestions to start replacing your judgement. That’s why you already have a healthy instinct to verify, to push back, to tell me to stop. Those boundaries are what keep your voice intact.
QED.
(PS. The cover image is AI-generated. Had to be done!)