Note to readers: This article contains content created by LLM ChatGPT. The story portion is the product of my prompt and ChatGPT’s response. The rest of the article is my unadulterated creation, thunk up out of my own brain. Please, contact me with questions related to this article or the series of articles.Last time, for the first time, I proposed a look at artificial intelligence. I suggested that the key to any threat it represents is how we use it. You remember? The innocuous tool that we can choose to use, like a pipe wrench, to bludgeon our enemies like the mindless zombies that they are. Of course, if we discover they are not mindless, or zombies, we must refrain from murder.

Last time, for the first time, I proposed a look at artificial intelligence. I suggested that the key to any threat it represents is how we use it. You remember? The innocuous tool that we can choose to use, like a pipe wrench, to bludgeon our enemies like the mindless zombies that they are. Of course, if we discover they are not mindless, or zombies, we must refrain from murder.

In week two of this series, I bring you the second scene of a short story written by ChatGPT. I will analyze its value as a fictional narrative in an unsophisticated manner.

Does ChatGPT work as a storyteller, raconteur, or teller of tales? As we answer the question, we can learn about the craft and skill of writing fiction.

How are we handling our fear of the Large Language Model Artificial Intelligence? 

Previously, I agreed that we should fear AI. But not its capacity to supplant the human creative mind. Our fear should be that AI will make us lazy as creatives, and superficial in our expectations of quality fiction. Don’t just switch it on like an old television from the ’50s. Use it with, oh, I don’t know, intelligence?

Remember, most of all, this exercise is entertainment that prepares us for a brave new world. (There, I’ve made that dystopian worldview reference once more. Repetition is the key to taking over your thoughts. Muah ha ha.)

The prompt continues.

ChatGPT Presents “Shadows in the Smoke”

Scene Two

Harland leaned back in his chair, letting the silence stretch out. He’d learned long ago that people tended to fill in the gaps when left to their own devices. Sure enough, the woman continued, her words spilling out like a confession.

“My husband… he’s involved in something. Something dangerous. I don’t know the details, but he’s been secretive, more than usual. And then there’s this woman… I think he’s been seeing her.”

The corners of Harland’s mouth twitched into a frown. He’d heard variations of this tale too many times, but it was the undercurrent of fear in her voice that kept him listening.

“Do you have proof?” he asked, his tone professional, detached.

“Just suspicions. A few late-night phone calls, unexplained absences… and then I found this.” There was a rustling sound, as though she were reaching for something. A moment later, she spoke again, her voice trembling. “It’s a photograph. Of them together.”

Harland’s mind raced. An affair was the most obvious conclusion, but he couldn’t ignore the possibility that something more sinister was at play. This city had a way of warping even the simplest of betrayals into something dark and twisted.

“Where did you find the photograph?” he asked.

“In his jacket pocket,” she replied. “He must have forgotten it was there. I—I didn’t know what to do. I thought about confronting him, but…”

“But you’re afraid of what he might do,” Harland finished for her. It wasn’t a question.

“Yes,” she whispered.

Harland glanced at the clock on his desk. It was late, and Eleanor would be waiting for him at home. He’d promised her he wouldn’t take any more cases like this, cases that kept him out all night, that brought danger to their doorstep. But the tremor in the woman’s voice hooked him, pulling him in deeper.

“Tell you what,” he said, reaching for his coat. “I’ll take a look into it. Discreetly. But if things start getting too heated, you need to get out, understand? This city doesn’t play fair.”

“Thank you, Mr. Harland. I don’t know what I would’ve done without your help.”

“Don’t thank me yet,” he muttered, hanging up the phone.

As he stepped out into the night, the cold air biting at his skin, he couldn’t shake the feeling that he was walking into a trap. But he’d committed now. He would see it through, for her sake—and for his own.

The city streets were quiet, the kind of quiet that only came when everyone else was too smart to be out. Harland drove through the empty avenues, his mind racing. A husband, a photograph, and a woman who was afraid to confront the truth. It was a familiar story, but there were too many pieces that didn’t fit together neatly.

He pulled up outside the address the woman had given him—a modest house in a quiet neighborhood, the kind of place where people kept their secrets buried deep. Harland stayed in the car for a moment, studying the place. The lights were off, and the curtains drawn. No signs of life, no hint of the turmoil inside.

Taking a deep breath, he stepped out of the car and approached the front door. His knuckles rapped against the wood, the sound echoing through the still night.

He waited, listening for any movement inside. But there was nothing. He was about to knock again when the door creaked open, revealing a woman standing in the shadows. Her eyes were wide, her face pale.

“You’re Harland?” she asked, her voice barely more than a whisper.

“I am,” he replied. “Can I come in?”

She hesitated for a moment before stepping aside, allowing him to enter. The house was dark, with only the faintest glow of a lamp in the corner. The woman led him into the living room, where she handed him the photograph with trembling hands.

Harland took it, studying the image under the dim light. It was grainy, the kind of picture that could’ve been taken with a cheap camera. But the faces were clear enough—her husband, arm in arm with a woman who wasn’t her. They were standing outside a building Harland didn’t recognize, the kind of place that didn’t belong in a respectable neighborhood.

“Do you know where this was taken?” he asked.

“No,” she replied, wringing her hands. “I’ve never seen that place before. I don’t even know who she is.”

Harland nodded, slipping the photograph into his coat pocket. “I’ll find out. But remember what I said—if things start getting dangerous, you get out.”

She nodded, her eyes wide with fear. “I will. Thank you, Mr. Harland.”

He left her standing in the doorway, her shadow stretching across the floor. As he stepped back into the night, he couldn’t shake the feeling that he was being watched, that something was lurking just out of sight.

He lit a cigarette, drawing in the smoke as he considered his next move. The photograph was a clue, but it wasn’t enough. He needed more, and he knew just where to start looking.

There was a bar on the other side of town, the kind of place where people went when they didn’t want to be found. Harland had a contact there, a man who owed him a favor. If anyone knew where that photograph had been taken, it would be him.

But as he drove towards the bar, his mind kept drifting back to Eleanor. She was at home, waiting for him, trusting him. He’d promised her he wouldn’t get involved in anything dangerous, but here he was, chasing shadows in the night.

He’d have to finish this quickly, for her sake. But deep down, he knew that some things couldn’t be rushed. And this case—this case was going to drag him deeper than he’d ever gone before.

What conclusions developed regarding ChatGPT’s capability to weave a compelling narrative?

The descriptive elements of a noir mystery continue as we read ChatGPT’s story. Consider Harland’s image while he is “letting the silence stretch out” so his client can fill the gaps. A common picture of a weary gumshoe up against the hardened creatures of the everyday grind.

Cliché ideas carry us onward when our hesitant witness reveals partial details, secrets, and undefined dangers. Of course, there is the other woman. Harland has heard variations on this theme before, as have we.

Proof? he asks, in the detached manner of his species. More vagueness presented with “a few late-night phone calls, unexplained absences” as evidence along with, ah, the condemning photograph. This gets us wondering the way Harland wonders. Who might it be? As these stories go, it will either be someone nefarious, or an innocuous bystander.

Of course, the real danger lies in his client’s fear of confronting her husband. We presume the man is so dangerous and unpredictable that approaching him with questions about fidelity would put her at severe risk. Harland compares him to the inequity of the city itself. Which leads that same Harland, ever the wise sleuth, doles out the greatest advice ever given to a woman under threat of a questionable husband… IF it gets WORSE, leave!

It bothers me a little that a supposed expert in human nature, especially its darker inner workings, would advise such a bold strategy for a naïve bride against her unprincipled husband. To Harland’s credit, it is the ‘50s and spousal abuse is poorly understood and not well defined.

Oblivious, the woman thanks him for thinking of her welfare. Harland, ever the classic private eye, reminds her, “Don’t thank me yet.” It’s sound advice that even Sam Spade would envy.

Harland drives across town to meet the new client. His intuition suggests it’s a trap. But he presses on because he’s “committed… for her sake… and his honor.” No matter the threat to life and limb, or to his professional liability. No matter the promise to his patient and long-suffering wife. The former create suspense. The latter makes us question his integrity. It raises the question whether adding a spouse to our private eye’s characterization overtaxes ChatGPT’s capacity to put together a believable storyline.

When Harland arrives at the home of the curious client, we find she lives in a “modest house in a quiet neighborhood… [where they keep their] secrets buried deep.” ChatGPT keeps the mood consistent with these telling phrases. Its use of key details like the neighborhood with “the kind of quiet that only came when everyone else was too smart to be out,” and a darkened house with “curtains drawn” while inside “only the faintest glow of a lamp in the corner” creates the atmosphere mysterious.

The woman greets him tentatively, allows him inside, shows him the photograph “taken with a cheap camera,” the faces just clear enough to reveal her husband with a strange woman in front of a sketchy building neither of them recognizes. So, Harland makes the promise, the promise of a good detective, “I’ll find her.” Because he knows, and the reader knows, her identity will unravel the mystery and define the threat.

Our detective leaves her at the end of an elongated shadow. Good imagery. A metaphor that might hold a clue. Or a red herring. One that makes us wonder if the “author” is pulling a fast one on us.

Harland “couldn’t shake the feeling that he” was being watched by a lurker. So, he lights a cigarette… to show he’s unafraid or unaware. To help him think. That’s what cigarettes do in noir fiction. He visits a bar to meet the man with all the answers. The fellow in touch with the underbelly of the city. A good man to know. Not one you invite to a barbecue in your backyard.

As he drives, he thinks of his beloved Eleanor, alone, trusting that he is true to his word about staying away from dangerous situations. Since he cannot compromise the case, he compromises his promise to Eleanor by trying to solve it as soon as possible. But even he knows that a rushed case increases danger.

ChatGPT’s use of familiar key phrases establishes mood and story motion. It comes off trite. Is still a bit stale. But if you’re reading fast and are hungry for an easy dose of crime, it does the job. The uninspiring dialogue from a thousand detective stories makes for a quick read. A great advantage of using an LLM to write a fictional account.

If you recall, LLM AI gathers bits and bytes from all across the land of preserved Internet data, relevant to the requested prompt. It seeks to generate words, phrases and syntax that copy the style indicated by the user. This feat is accomplished in a manner similar to the process humans use to assimilate and regurgitate. Albeit, with a mite less emotion and a tad more mathematics.

These posts hope to explore the components of storytelling as we evaluate ChatGPT’s construction. The quest here is one of awareness and discernment. We seek to identify whether AI, in this case ChatGPT, can produce material of equal quality and value to that of a human mind.

Let me know your thoughts.