Last year, I wrote about ways generative AI can help writers without infringing on others’ copyright and our creative expression. Much has happened since then. Now, the creative community has gone into a full-blown rebellion against generative AI. Agents, publishers, authors, and readers are busy sussing out writing that is AI-generated. They look for signs of ChatGPT in every em dash and out-of-fashion term. Some link generative AI with fascism.

My own perspective about generative AI for writing has also changed.

I discontinued using an AI-based tool for editing. Part of my decision came from my frustrations with the tool itself. I wasn’t always happy with its editing recommendations. I often repeat the same opening phrase in multiple sentences for emphasis, which is a carryover from my Toastmasters experience. But the tool flagged the repeated phrase as a mistake. And I had character names flagged as misspellings, even if I added them to my dictionary. This isn’t only a problem for invented names. I have a character named Yusef. This is a perfectly normal Arabic name, but my spell checker still slaps it with a red underline.

What ultimately swayed my decision to discontinue the tool was the change in political environment. The tool vendor can pinky promise your content won’t be sucked into their language models, but your work is still under their surveillance. The software is analyzing your writing, processing it on their servers, and returning suggestions based on their algorithms. How do you know they’re not reporting on the things you’re writing? What if the changes they make undermine the messages you want to deliver? Considering that a number of these companies have close connections to the current regime, how can their software be trusted?

While expressing yourself is a risk these days, I want control over what I write and when and how I choose to make it public. I don’t want my software conspiring against me.

Since I dropped that tool, my writing has gotten better. I can express myself more freely. No more constant corrections that hampered my flow. I can focus on what I want to express instead of feeling I have to make those words perfect as I write them. A few typos is a small price for creative freedom.

The problem with generating books with tools like ChatGPT is your words are not your own. You become a mouthpiece of whatever the AI produces. Unless you’ve done the research and writing yourself, you have no idea if what was written is true and accurate, and if those words align with your own views. You surrendered your insight and creative expression to an algorithm.

I still believe generative AI has a place in writing. It’s useful in research. It can summarize ideas so you can use them as a starting point for further investigation. And I’ve found it very helpful in creating reference images to develop characters.

In the novel I started, my main character is a 29-year-old technical writer. (Write what you know, as they always say.) In one scene, she comes home from a frustrating day at work and a disappointing date. She changes into comfortable clothes to play her favorite vintage video game on the Commodore Amiga 500 she inherited from her father. This is what Microsoft Copilot came up with.

This is your stereotypically bad AI image. She faces forward, but the screen is on the side. The Amiga is a desktop computer. She would be sitting in a gaming chair, not the floor. And what’s with the t-shirt, Copilot? I couldn’t use this image for book covers and promotional art because it contains artwork and photography from others. At least she has the correct number of fingers.

But for developing character, this image is extremely useful. I’m able to develop her backstory and personality, create details about her daily life, and develop dialogue. I can imagine the perceptions the other characters would make from how they see her and use them to create conflict. I can keep details about her consistent. Entire scenes can be built on a visual prompt like this one. When used this way, generative AI sparks creativity instead of being an artificial end-product.

Generative AI, like all technologies, can be used for good or ill. It can stimulate creativity or stifle it. As writers, we shouldn’t impulsively reject or surrender to it. We need to understand it and learn how to use it properly to further our imagination and create better and more human writing.