Note to readers: This article contains content first created by LLM ChatGPT, and later modified by yours truly. The story portion began as the product of my prompt to ChatGPT and ChatGPT’s response. A rewrite of the story follows the introduction and was completed by myself, dredged from the depths of my personal organic database. Please, feel free to contact me with questions related to the content of this article or the series of articles.

In a rapidly changing technological landscape, artificial intelligence poses a threat for many, an opportunity for others. After the series of articles detailing our interesting romp across the landscape of a ChatGPT generated story, I’ve decided to give the narrative a once over to fix the parts that didn’t work. 

In the upcoming weeks, I will share my corrections to the short story written by ChatGPT. Each episode will include a separate scene, as was the case previously. In that collection of posts, my intent was to evaluate how well ChatGPT worked as a storyteller, raconteur, or teller of tales.

This round, I apply the human element with some notes at the end of each scene describing my thoughts on the changes, as well as on the craft and skill of writing fiction.

Are you up for the journey? Let’s get after it, then.

I’ll tell you a story, scene one.

Shadows in the Smoke

This city held secrets. Down dimly lit alleys, tucked into the corners of the smoke-filled bars, within the walls of the shiny bits of glitz and glamour, a labyrinth of unspoken deceits thrived in the shadows.

John Harland knew his way around this maze. He understood the twist and turn of it, which darkened doors to avoid, and where to stand to hear a whispered lie. 

In his years since becoming a private eye, Harland had observed a growing darkness that swallowed the light—cheating spouses, stolen goods, missing persons who refused to be found. 

Tonight he sat behind the wheel of his ‘46 Buick, staring at a worn photograph he kept tucked in the visor. The bright smile aglow with youthful energy, eyes peering straight into his soul, while she swung gaily from a pole by a single, graceful arm. Before the tough ones, the cases that gnawed at the lining of his stomach and made him queasy, he’d pull out the photo and gaze into it. She grounded him, this gorgeous and tolerant lady— pulled him back from the edge of dread. If he stared into the photograph long enough he could see the moment come to life again. Their first days together, when he was galant and fearless, before marriage and kids and obligations weighted their balloon with responsibility. 

“Eleanor,” he said in hushed reverence. “My gut tells me to go home. Sit with you in the safety of our blessed home.” These were words he dare not share in her presence. She would tell him to end it, retire, settle down  like they’d planned.

He flicked his cigarette into the street, watching the ember die out in the rain-soaked gutter. It was the final stick from the hidden pack he’d purchased during his last case. If he didn’t quit them soon it would be his ember bleeding out in some dark boulevard. 

The client had been vague, just enough intrigue to pull him in. But also, there had been a tremor, a hesitation in her voice that made him uneasy. Again, his gut told him that this was a job he should leave alone. Only it didn’t make sense to run. He had no identifiable elements that suggested abnormal danger. 

Besides, the bills didn’t pay themselves. This client smelled like money, and promised a good bit if he helped her. Eleanor waited at home, patient and loyal, certain that his skills would protect him. Convinced that he would bring home the proverbial bacon until the day he retired. Then they would move to a quiet neighborhood with civil dogs and friendly mailmen. Just this one more job.

Private eye work was unpredictable. Eleanor was constant. Between them they’d survive it. Like they always had. One final job, one final payday, and he’d hang up his fedora for good. For Eleanor. 

He climbed out into the wet night, wound his way up the three stories to his office, picked up the phone and dialed the number scrawled on the back of a matchbook. The line rang once, twice, before a woman’s voice answered. Soft, almost too soft to hear.

“Yes?”

“Harland,” he said, voice steady. “You still need that help?”

A long pause punched a hole in Harland’s confidence. For a moment, he thought she hung up. When she spoke again, her voice was thick with ambivalence.

“Yes,” she whispered. “But it’s more complicated than I thought.”

What did I do to make it human?

Although this story opens with strong imagery, those images sounded flat and uninspired in the hands of ChatGPT. My idea was to tune them up a bit by aligning the phrases more with John Harland’s voice. I also recognized how disconnected the imagery was from a realistic flow. For example, John was sitting in his car and then suddenly on the phone with no transition. This wasn’t possible in the 1950s. 

I also wanted to incorporate the split personality of a jaded John Harland, P.I., juxtaposed against the dedicated husband who needs to make a buck but understands the threatening nature of his work. 

Details like “behind the wheel of his ‘46 Buick,” and “staring at the worn photograph” were not bad in themselves, but I chose to create a little more distance between John’s profession and his home life. Noir detectives don’t typically have spouses to complicate their cynicism, so this element will be interesting to work with. I felt that ChatGPT’s version lacked the nuance required to make this separation between hardboiled P.I. and loving family man believable. 

In the original, Harland’s claim that this was his “final job” so he can “hang up his fedora for good” sounded trite. I tightened the narrative flow to pull out a genuine and true feel for the character. 

The dialogue appeared uninspired upon my first analysis, echoing the dialogue of a thousand detective stories before it. But I discovered that a simple change in assumptions allowed it to work effectively as a wrap to the scene. Once I took out the woman’s assumptive response when she picked up the phone, that Harland was returning her call— a dangerous oversight for both the woman and the writer—  the simplicity of the dialogue became powerful enough to carry the key question: How dangerous is this case?

That’s a highlight of my reasoning for scene one changes. 

My hope is to work through this story quickly, over the next couple of months. Each scene will present its own problems, which might stretch the process more than I anticipate at this point in the process. I ask you to bear with me. Let’s see what value we can find within the molded plastic bones ChatGPT has provided.

Let me know your thoughts.