In an effort to explore new creative avenues in my retirement, I enrolled in a community education poetry class. For our first assignment, we met at the Boca Raton Museum of Art, where the task was to write a poem about a work that inspired us. I chose to focus on a blown glass piece by Liesa Cole. She had spelled the word “TRUTH” backwards. Intriguing.
I wanted to play with the idea of two sides to every story, and how perception can be reality regardless of the actual truth. But as I began to write, I was paralyzed by the negative, nagging voice in my head telling me my words were not descriptive enough, my expression not creative enough. I found myself, for the first time, facing the dilemma that students face daily: what would ChatGPT write if prompted?
There is a peculiar irony in this struggle, and my dilemma is deeply personal. I’m a teacher. I spent my career impressing upon my students the value of doing the work oneself, and the seriousness of academic integrity. Using AI feels like a shortcut, a way to bypass the very struggle that gives writing its meaning.
And yet, ignoring a tool that is now so widely used also feels like resisting our present reality. On the one hand, it is a tool, much like dictionaries, thesauruses, and style guides that I’ve relied on, in all of their evolving forms, since high school. On the other hand, it challenges my sense of integrity. If I lean on AI to generate sentences, to structure my arguments, or even to suggest phrases, am I still writing?
Every age grapples with new technologies and their ethical implications. Printing presses, typewriters, and word processors. Calculators, computers, steroids. Autotune, autocorrect, autopilot.
Cloning, stem cells, nukes. Algorithms.
The challenge is knowing where to draw the line between convenience and conscience.
The irony of using artificial intelligence to help craft a poem about truth is itself poetic. The tension between convenience and integrity is the very subject of my poem, turning my dilemma into part of the creative process.
Ultimately, my poem conveys my idea. I entered a prompt into ChatGPT, and it spit out 300 words in less than 10 seconds. But it wasn’t quite what I wanted to say. I re-prompted, and it rephrased the entire paragraph in a way that better reflected my meaning.
Again, in less than 10 seconds. After numerous attempts at recrafting the narrative, I had enough material to cherry-pick the best parts and weave in my original words and sentences, creating a tapestry of my ideas using both original and manufactured sentences tailored by me to reflect my meaning. I guess the question is, did I cheat?