3 Ways I’m Using AI Ethically as a Novelist (Spoiler: It’s NOT to Write)
Here’s the transcript of the podcast episode:
How I Use AI as a Writer
Today, I want to share the ways I'm using AI as a writer. I'll share three specific applications that have genuinely enhanced my writing process without dipping into ethically or professionally questionable territory.
I want to say from the top, because this is a controversial subject, that I don’t use AI to write—as in, I don’t have A.I. do the writing.
And that choice is not even primarily for ethical reasons. It just isn’t a good enough writer. It’s arguably better than some people who aren’t strong writers…maybe.
But if you’re a writer by profession or write a lot as part of your work? You’re almost certainly better than AI. And even if it could be 80% as good as me…80% as good as me isn’t good enough, you know? Not when the writing is the job.
Also, writing is the fun part of being a writer.
It’s like my sister said the other day, "Can we teach AI how to do our laundry and wash our dishes so that we can do the creative stuff?” Maybe let’s not train the bots to do the best part of our jobs?
But just because I’m not using it for writing doesn't mean that it isn’t helpful for other things. I’m going to tell you about 3 things I find it helpful for—the first is for research, and the second and third are a continuation of the last episode, where I talked about how I dictated much of my most recent novel.
Using AI for Multi-Dimensional Research for My Novel
First, my favorite way to use AI for novel-writing is for multi-dimensional and nuanced research. I view it as a kind of Google 2.0. It can handle complicated search queries that would otherwise require multiple separate searches and cross-referencing.
For example, in writing a scene for my new novel, I needed to find a pink gemstone that isn’t super expensive and has a fun history, because the person selling it needed to be able to tell a story about it.
In the past, pre-GPT, I’d google “pink gemstones,” then find ones I like and check their prices, then google the names of those to search for any lore or fun associations. This would be a multi-step process that would take me at least fifteen minutes, probably more like thirty or forty-five.
But now, I just plug my multi-dimensional request into Claude, which is my AI of choice because of its ethical standards and data use policies:
I need a pink gemstone that one could buy in a piece of jewelry for a few hundred dollars, and I want it to have fun historical facts associated with it.
Within seconds, I have a great option: the pink spinel, a gemstone I’d never heard of.
Here’s another example. For the same novel, I typed this into Claude:
I need to find a medical condition that can cause single eye blindness. It can go undetected at birth and is treatable. I'd like for it to be possibly caused by environmental conditions or genetically passed down (I'm not saying it can be either; I am saying I want a condition that COULD be either in a given instance).
AI found me one really quickly.
I want to be clear that I'm not wholesale taking AI's word for it that whatever it tells me is true. We know it hallucinates.
I view this kind of research as a starting point, not an endpoint. It's shortcutting the BEGINNING of your research—but it’s a great beginning. Because now I have a strong possibility that could work, I just need to confirm it.
The second and third uses I want to tell you about both involve dictation.
Novel Dictation + AI: Cleaning Up Grammar Without Losing Your Voice
As I mentioned in the last episode on how I’m using dictation to write my novel, I've been dictating portions of my novel while walking. If you want to learn more about how I’m doing that, go back and listen to or watch that episode, but in this episode I want to talk about how I use AI to help me sort out what I’ve dictated.
There are 2 categories of dictation that I used in working on my book: dictation of actual scenes, and dictation of brainstorming.
Using AI to Clean Up Dictated Scenes for My Novel
The first category of my dictation is actual writing for the book. I set out with a particular scene in mind, go on a walk, and dictate until I finish that scene, which takes anywhere from 45 to 90 minutes. I do this in the Notes app on my phone, not worrying too much about grammar or naming conventions.
As soon as I get home, I need to do something with it or I'll get confused later about what writing is where. So I copy and paste from the Notes app into my AI of choice, which, again, is Claude. (I’m not getting a kickback from plugging Claude here by the way.)
I paste my dictated text with a very specific instruction: I say “clean up the grammar but don't change any of the words themselves.”
This last part of the instruction is very important—I don't want AI writing anything for me from a craft perspective, and it will try to change things unless you explicitly tell it not to.
So I always instruct it explicitly not to change any of my words, only to clean up the grammar.
In general, grammar cleanup is something it’s really good at. It's sort of like a curmudgeonly eighth-grade teacher who loves to nerd out on commas. If you heard or watched the previous episode on dictation, you’ll know that for conversations, I’m not dictating quotation marks. Claude will get some of these right and some of these wrong, so I have to do a clean up of its clean up, but it’s still a huge time saver.
Using AI to Summarize Dictated Brainstorming
The other category of my dictating is brainstorming, and this is the third and final use for AI that I’ll be talking about. To be clear, just like in the instance above, I’m not using AI while I’m dictating. I use it after I have dictated.
When I’m using dictation to brainstorm, I’m really just thinking out loud, almost like talking to a friend. And this is where dictating really shines for me in terms of story development. I love it because what I’ve found is that in talking out loud to myself as I'm walking, I can have a lot of really wild and cool ideas.
I use the native Voice Memos app on my phone. I just start recording myself talking out ideas. Sometimes I talk for 20 minutes, sometimes for 45 minutes, but I just talk through any questions I have about the plot or what's going to come next. I'm not transcribing this in real time; I'm just recording it .
When I get home, I transcribe it. You can transcribe directly in the Voice Memos app, but that functionality is a little glitchy. I have better luck downloading the audio file onto my computer and transcribing it using other software. I use Descript, which is a paid subscription, but I’m sure there are free ones out there.
This transcription is way more of a mess than my dictated scenes are, because in brainstorming sessions, I’m not even trying to be coherent. I have so many false starts, so many ums, so many meandering thoughts that lead to nowhere. Trying to follow my train of thought is like chasing a toddler around a room.
And this is where AI comes in handy.
I drop the transcript into AI, but this time I tell AI to “pull out the decisions I landed on and summarize my major takeaways.” It’s really great at this! It’s like a very smart intern who got As on all her book reports.
So if I spent 45 minutes just talking out loud about what I think maybe could happen next in the novel, and I had seven false starts for ideas I discarded but landed on one I really like in minute 37 and then spent the next 8 minutes developing it, AI is really good at figuring that out from what I’ve recorded myself saying.
It’ll churn out the takeaways much more efficiently than me scrolling through that entire transcript looking for them.
Now, is it perfect? You know it isn’t. But it still saves me time, and that’s time I can then spend writing.
The Three Ethical AI Applications - Summary
So there you have it—three ethical ways I use AI in my writing process. To summarize, I use it:
For multi-dimensional and nuanced research that would otherwise require multiple searches
For cleaning up grammar in my dictated scenes without changing my actual words
For synthesizing the key points from my brainstorming sessions
These applications help me work more efficiently while maintaining complete creative control. AI isn’t writing for me—it's handling the tedious tasks so I can focus on the creative part.
Thanks for listening, and I'll catch you on the next episode all about what I learned from Ann Patchett!
How much money can you really earn as an author?
I gathered data from more than 1,400 published authors to dig into the realities of book advances, self-publishing income, and the key factors that affect them, like genre, debut status, agent representation, and publisher type.
Dig into all this information with me in a free workshop filled with real-world insights, practical strategies, and visual tools to help you navigate your publishing journey with confidence.
If you’re determined to understand your earning potential and make informed decisions for your writing career, this workshop is exactly what you need.
Come say hi on Instagram!
