AI Seduced Me
One man's journey with generative AI
I was seduced.
I used AI for the first time about four years ago. I was managing a marketing team of five people for a manufacturing firm. AI was new and I logged in and played with it having it suggest edits for some social media copy and create some clever CTAs. It was harmless enough. We didn’t use it all the time, but occasionally when the creative well seemed dry, we leaned into it—made some images, adjusted a few words.
I only glimpsed the issue then.
After I retired, I had little use for AI, until one day, a little over a year ago. I had an idea for a story, a fable of sorts. I wrote the first chapter and thought, “I bet AI could help me develop this.”
I shared the piece with ChatGPT and it took off rewriting it, asking me questions about what would happen next and in the span of minutes, my responses and its writing generated a version of my fable that read–to me at the time–marvelously better than anything I could have done alone. I shared it with a few people and they loved it. I published it here.
I created a few more pieces that way. It felt simple, easy and a bit miraculous. The pieces were well received, it appeared. My ideas were getting out there and I was avoiding the usual plagues of typos and poorly constructed sentences.
Then, something shifted. It happened one Wednesday afternoon when I was working on a sermon for that week. I took the rough draft over to AI and asked for help. How can I describe this? I fed my message into AI and what came back was startling. AI removed my unique sermon style, whitewashed the tone and even tilted the theology in a different direction. I realized—I wouldn’t preach that message. When I questioned AI about that, it treated me with a somewhat patronizing tone. Although it sounded like it agreed with me—even offered some revisions—it was arguing for its specific take on the message. I stopped.
And it hit me; I wouldn’t have written those Substack essays like that either. Yes, I began to see the patterns of writing, the AI rhythm, the overused phrases, BUT what really hit me was that I had allowed it to replace my own opinions and voice.
I didn’t swear off AI. I figured I must be just using it wrong. I began to read articles about creating prompts, implementing “skills” and training AI to a new voice. I created personas with specific guidelines. I tried skills. But the models were changing so fast that over the course of a few months I experienced increased resistance to my efforts.
Finally, I gave up.
I still use AI (especially non-generative forms) to help me with my own typing and writing issues. I have used generative AI to help me with mechanical work on the novel I’m writing and to spot some logic gaps and timeline issues in that work, but I refuse to let it write for me. It still tries. AI will suggest a correction to something mechanically wrong, with a phrase or sentence that is classic AI speak. Often that phrase carries a different meaning or intent than my own.
My journey continues. I’m reading more now about the ecological damage that LLM technology, specifically for generative AI, is causing. I’m listening to poets and novelist as they present a case for and against the use of AI. I’m watching as the creators of our technology increasingly lean into our tech to give us what they think we want. I’m not sure where we are going, nor am I clear on what we should do in response to these challenges.
I do know this. Generative AI isn’t neutral. AI has opinions and will readily apply those to your creative process—invited or not.
Below this rude paywall is a three day Growth Guide for paid subscribers. It walks you through an exploration of your use of and perspective about AI.
