Readers are advised: this is fairly long because I include examples of my “conversations” with ChatGPT (about 4,800 words). Prior posts about ChatGPT are HERE and HERE and HERE.
Introduction
In the last ChatGPT post, I addressed ChatGPT’s function as a source of information. In this post, I’ll address the creative aspects of ChatGPT.
By now, readers have been exposed to many articles about ChatGPT. Pro, con, neutral, alarmist, enthusiastic . . . there is no shortage of opinions, so what makes me think I have anything to add to the conversation?
Well, not to the conversation, per se. But I can add what ChatGPT might mean to me and my writing, how I might use ChatGPT in my fiction and non-fiction writing, and what I think about it permeating into every facet of what I interact with.
Let me give the spoilers for the many who are not prone to enjoying the meandering of my thought. Not much, I won’t, and I’m simultaneously indifferent and mildly troubled.
OK, with that out of the way — and for the few who will keep reading — let the meandering commence!
AIs as creative tools.
Readers might notice these posts make use of AI-generated images. Intrepid readers familiar with my opinion posts will recognize this as a departure from the norm.
I usually adorned my opinion pieces with photos. Specifically, my photos. That practice started when I recognized that many people are turned off when faced with a page of text. Such pages come across as chores, something one has to struggle through even if mildly interested in the subjects of my rants reasoned opinions. The photos break up the offering into palatable morsels.
And yet, I was never entirely comfortable with the practice, even as I told myself it was for the benefit of non-readers. The problem was that I used photos I might have otherwise featured in photography-centric posts.
These AI-generated images, on the other hand, occupy a niche that I’m comfortable with; they mean little to me since I didn’t generate them, they’re visually interesting, and while not of my making, they are at least partially the result of my efforts (thinking up the prompt and coupling it with one of my photos).
Heck, I don’t even care if someone steals them. Well, maybe a little, but nowhere near how much I care about my photos.
So, this is an example of AI indirectly being used as a tool to help my creative efforts. It contributed something I wasn’t willing or capable of generating on my own.
Would I do the converse? For example, would I use ChatGPT to generate words for a post showcasing some of my photos?
No.
You might think it’s the same, but it’s not. Let me explain.
The distinction is this: generating graphic images is not something I ‘do’. . . I don’t generate art from scratch. I don’t draw (although I would like to) and don’t use my meager computer skills to create computer graphics.
These AI images aren’t competing with anything I might do; they don’t compete with the few skills I have.
Having ChatGPT write something for a post, on the other hand, competes directly with what I enjoy doing: writing. Some have even said I’m not bad at it (and I thank them for it).
But isn’t ChatGPT just a tool?
Indeed, there is no shortage of writers expressing the opinion ChatGPT is no more than a tool in an arsenal of tools.
I disagree.
When I write something, it’s equivalent to me snapping a picture. Meaning I initiate what is to be the final product: my idea, my vision, and my execution.
The tools I might use, for example, Lightroom for photos and Grammarly for writing, are just that, tools. They are used to shape the final product, but that product came to be by my hand. Those tools have no function absent something to apply them to, like my photo or a piece of my writing.
“But, Emilio, AIs generate images and paragraphs based on your direction and input! How are they not tools?”
Hmm . . . I see the problem, but I’m unsure if I can define a boundary between what I consider a tool you do stuff with versus a tool that does stuff for you. The problem is that the differentiation can appear as a gray area to some, but for me, it’s as clear as the difference between night and day.
But, let me do what I always do . . . use an example.
Say I want to build a flowerbed. I could buy lumber, bring in dirt, and buy various construction tools . . . or, I can hire a company to come in and build it per my directions.
I could conceivably claim the end product is my design, but I couldn’t claim that I built it. And even then, the end product might not be exactly what I envisioned. I say this from previous experience with hiring people.
See the difference? No?
How about this: I want to do a sculpture for my yard, so I buy a marble statue of a hummingbird, chip a piece of marble from it, and slap a plaque on the statue saying that I made it . . . how many would agree the plaque was truthful?
“But, Emilio, doesn’t it depend on how much you modify the statue?”
Well, Bob, I suppose you can argue for some credit, but where do you draw the line?
Plus, I still see a bigly difference between starting with a giant cube of marble versus starting with something already resembling the end product. For one thing, adding marble to a piece already shaped is difficult; you are restricted in your creative input to what’s already there.
“But, Emilio, writing isn’t a piece of marble. It’s easy to modify and even add words!”
Well, Bob, if you are a writer, you know the difficulties and limitations of editing an existing piece of writing. Even if it’s words and not marble, you’re still constrained by what’s already there and often have to start from scratch if you want something different.
Look, Bob, I know you have a basis for an argument, but at the core of the discussion is whether you are the originator of the piece.
Is having an idea sufficient to claim ownership of the end product?
Perhaps for some, in which case they should do what some people do . . . use ghostwriters and slap their names on the finished product.
Let me summarize it thus: writers write books, and editors suggest changes, not the other way around. I write the story, and Grammarly tells me suggests corrections to the use of commas, verb tenses, etc.
How about using ChatGPT to originate ideas?
The current Title Writing Challenge is based on the writers’ interpretation of the title “The Invisible Man”.
After I had an idea of what I wanted to write, I was curious. What would ChatGPT do with the challenge?
Chat GPT Plot Request 1
How about an idea for a short story entitled “The Invisible Man” but without actual invisibility?
“The Invisible Man” tells the story of a man named John who has always felt invisible in his own life. Despite his professional success and a loving family, he has never felt truly seen or understood by those around him. One day, while on a business trip, John discovers a powerful new drug that makes him physically invisible to others. At first, he revels in the freedom and power that invisibility brings him, but soon realizes that it also means he can never truly connect with anyone again. As he struggles with the isolation and loneliness of his newfound invisibility, John is forced to confront the truth about himself and the people in his life, ultimately learning that true connection and understanding come from within.
I mean “invisible man” as a metaphor, not as an actual ability.
“The Invisible Man” tells the story of a man named David, who is a successful businessman and a loving husband and father. Despite his achievements, David feels unfulfilled and disconnected from the world around him. He’s always felt like he’s been living his life on autopilot, going through the motions and never truly being seen or heard.
One day, David decides to take a sabbatical from his job and sets out on a journey of self-discovery. He travels to different parts of the world and meets a variety of people who teach him to open up and connect with others on a deeper level.
As David’s journey progresses, he starts to realize that the invisibility he’s been feeling is not due to any external factors, but rather his own self-imposed isolation. He starts to see that the key to feeling seen and heard is to truly see and hear others, and to be vulnerable with them.
In the end, David returns home, a changed man. He’s more present in his relationships, more engaged with the world around him, and most importantly, he feels like he’s finally found his place in the world.
. . . how about people we take for granted but don’t notice.
“The Invisible People” is a story about a man named Jack who is a successful and busy executive. He’s always on the go, focused on his career and his family. He doesn’t have much time to think about anything else.
One day, while walking through his neighborhood, Jack sees a group of homeless people huddled on the side of the street. He’s never really paid attention to them before, but something about them catches his eye.
Jack starts to notice the people he’s been taking for granted all around him – the janitor at his office, the waitress at his favorite restaurant, the barista at his local coffee shop. He starts to realize that everyone has a story, and that the people he’s been ignoring are just as important as he is.
Feeling guilty and remorseful, Jack starts to volunteer at a homeless shelter, and he begins to understand the struggles of those who are invisible to society. He starts to see the beauty in the world that he’s been missing and starts to enjoy the small things in life again.
Jack’s journey of self-discovery teaches him that everyone is important and that we should never take anyone for granted. He becomes a more empathetic, compassionate, and caring person, and starts to see the world in a whole new light.
I sense a common theme of self-discovery, but I’d prefer something more like someone who is very influential even as he goes unnoticed through his daily life.
“The Unseen Leader” is a story about a man named Michael who is a highly successful and respected business leader. Despite his accomplishments, Michael is a humble and unassuming person who always puts others first. He goes about his daily life with little fanfare, and most people have no idea of the positive impact he has on the world around him.
Michael’s leadership style is based on empowering others to be their best selves, and he’s always been more comfortable working behind the scenes. He’s the type of leader who doesn’t need to be in the spotlight to make a difference.
One day, Michael’s company is going through a difficult time, and the employees are feeling unmotivated and uninspired. They’re ready to give up, but Michael is able to rally them to keep going, with a quiet and determined spirit.
As the company starts to turn around, Michael’s colleagues begin to take notice of the unseen leader who has been guiding them all along. They come to realize that his leadership and influence have been the driving force behind their success.
The story illustrates the idea that true leadership doesn’t always come in the form of a charismatic figurehead, and that sometimes the most powerful leaders are the ones who work quietly and behind the scenes. It also shows that people who are not seeking fame or attention can be great influencers and leaders in society and that these individuals can make a significant impact in the world.
I’ve read these types of stories before — much like the AI probably did — and while I briefly considered this kind of theme, I dismissed them out of hand as none lent themselves to either humor or action (my preferences when writing). One should also note that the writing isn’t particularly good or even grammatically correct in places.
I asked ChatGPT if it could suggest something like what I might write.
Chat GPT Plot Request 2
Can you write it as Emilio D’Alise might write it?
“The Unseen Leader” by Emilio D’Alise
In the boardrooms and executive suites, a man walked with the grace of a lion and the humility of a lamb. He was Michael, a leader among leaders, a man of great power and influence, yet he went about his daily life unnoticed, like a ghost among mortals.
But little did the world know, that this man, this quiet and unassuming figure, was the driving force behind the success of one of the most powerful companies in the world. He was the unseen leader, the one who worked tirelessly behind the scenes, pulling the strings and guiding his team to greatness.
The company was going through a crisis, the employees were feeling unmotivated and uninspired, they were ready to give up, but Michael, with his unwavering determination and his silent strength, rallied them to keep going.
As the company started to turn around, Michael’s colleagues began to take notice of the unseen leader who had been guiding them all along. They came to realize that his leadership and influence had been the driving force behind their success.
In the end, it was not the accolades or the fame that mattered, but the impact that one individual can make on the world, when they lead with integrity, humility, and a true sense of purpose. Michael’s legacy will not be measured by the trophies on his shelf, but by the lives he changed for the better.
Well, obviously, that’s not what I had in mind. The AI just plopped my name as the author but didn’t change anything or reflected my style of writing.
Chat GPT Plot Request 3
What do you know about Emilio D’Alise and his blog?
I’m sorry, I do not have any information about an Emilio D’Alise or his blog. My training data only goes up until 2021 and I am not aware of any individual by that name who is a known blogger or author. It is possible that is a fictional person you are using in this context or a relatively unknown writer. If you could provide me with more details or context I could provide you with a more accurate answer.
His blog is DisperserTracks, a WordPress blog, but you are right. He’s, or rather, I’m, not famous. I was just wondering if you sampled my blog or if it’s too obscure for powerful AIs to bother with (not a bad thing, that).
I apologize for any confusion, as a language model AI, I don’t have the capability to browse the internet or access any external sources. My knowledge base is based on the data I was trained on, which is a large dataset of text from a variety of sources, but I do not have any knowledge of your blog specifically.
The reason I ask is that I know you can write in the style of people familiar to the public, and that seems a bit egalitarian, further minimizing the contribution of everyday people in favor of people who often are superficial and more focused on style than substance.
As a language model AI, my goal is to assist users with generating human-like text based on the prompts given to me, regardless of the fame or obscurity of the person or style being referenced. I can generate text in the style of well-known individuals, but that is because I have been trained on a large dataset of text from a variety of sources, including texts written by well-known individuals.
It’s important to note that my ability to generate text in a specific style is based on patterns and structures found in the text I was trained on and it does not reflect any bias or preference for any particular individual or group of people.
I understand that you may be concerned about the way in which the contributions of everyday people may be overlooked or minimized in comparison to those of more well-known figures. It’s true that in many contexts, the attention and recognition is often given to those who are more famous or have more visibility in the media, however, It’s not my intention to minimize the importance of everyday people, but rather to help users with generating text based on the prompts given to me.
And this is another crucial thing about using ChatGPT as a tool for my creative efforts; it would have to “know” me.
From what I hear, you can feed it samples of your writing and ask it to expand it, change it, improve it, etc. . . . something I did in one of my chats. I asked it to expand a Fifty Words Flash Fiction Challenge titled Wounded (LINK). Here is the piece as written:
Wounded
© E. J. D’Alise, 2015
”No disease, no drugs.”
“How much?” Volpe asked.
“$200 per pint.”
“I’ll take two,” Volpe said.
Mary let her blood flow, filling the container to the four cups mark as the vampire hungrily watched.
The wound in her arm was a godsend; Mary had many bills to pay.
And here’s what ChatGPT returned.
ChatGPT’s Expanded Story
“No disease, no drugs.”
“How much?” Volpe asked.
“$200 per pint.”
“I’ll take two,” Volpe said.
Mary let her blood flow, filling the container to the four cups mark as the vampire hungrily watched.
Volpe watched as the blood filled the container, his eyes bright with hunger. He had always been a bit of a thrill seeker, and the thought of drinking human blood had always appealed to him. He had heard about a underground market for it, and after a bit of digging, he had finally found a supplier.
“Thank you,” Volpe said, handing Mary the cash. “I’ll be back for more next week.”
Mary nodded, trying to ignore the feeling of disgust that welled up inside of her. She had never been comfortable with the idea of selling her blood, but she had no other choice. She needed the money, and this was the only way she could get it.
As she watched Volpe leave, Mary couldn’t help but wonder what would happen if she were to get caught. Selling blood was illegal, and she knew that if she were caught, she could go to prison. But she didn’t have any other options, and the thought of going back to living on the streets was even more frightening.
So Mary continued to sell her blood, hoping that one day she would find a way out of this life. But until then, she knew she had to do whatever it took to survive.
ChatGPT missed the bit about Volpe being a vampire, and I pointed that out . . . and ChatGPT just added the word to a revised version:
“Volpe watched as the blood filled the container, his eyes bright with hunger. As a vampire, he had always been a bit of a thrill seeker, and the thought of drinking human blood had always appealed to him. He had heard about a underground market for it, and after a bit of digging, he had finally found a supplier.”
By the way, given the blood shortage, I probably should have had the price as a more reasonable $1,000.
ChatGPT missed the story’s point, and what it added was not what I consider an improvement.
Perhaps it might do better with a larger sample or more direct instructions.
So, how might one use ChatGPT?
ChatGPT as an aid to learning to write
I’ve already said I wouldn’t use it in my writing (for reasons), and my limited experience with it tells me it has little to offer (me) beyond being a curiosity.
I can almost visualize someone using it as a training tool, but I’m skeptical of its usefulness.
Conceivably, it might not differ from someone reading many well-written books and identifying what makes them so. Some might even find general ideas like what ChatGPT offered above as jumping-off points.
Some writers claim it saves them time in writing descriptions of places, things, and even people.
Again, not something I would do, but I have to take them at their word . . . with a huge caveat.
If you’re a writer who’s already found a voice (HERE and HERE) and know what you want from your writing and how to achieve it, then ChatGPT might be a tool that helps speed up the process.
But if you are not a writer with hundreds of thousands of words under your belt and you start using ChatGPT to fill the gap . . . well, you’re doing yourself and your readers a disservice.
I’m firmly in the “ten-thousand hours of practice” camp for learning a skill. For writing, I think there’s a saying that “the first million words are practice.”
Just like I will never learn how to draw if I use DALL-E and/or MidJourney to generate drawings, I don’t believe you can learn how to write by having ChatGPT do the writing. All you’re learning is how to use ChatGPT.
Sadly, that might be enough for many.
Side note: Perhaps not equivalent, but look at what happened with the advent of calculators. Demonstrably, many people are incapable of doing basic math, be it in their heads or on paper. I believe that has a profound negative impact on their everyday lives. Others maintain calculators are just tools that freed humans to pursue other interests . . . like Candy Crush.
Remember, you can learn a lot from reading other writers, but doing it is different. You might start consciously or unconsciously copying the style of other writers, but the more you write, the more you’ll develop a style and voice unique to you.
I don’t think that can happen if you let ChatGPT do your writing.
. . . I could be wrong.
Side note 2: One ancillary skill to writing is doing so longhand in cursive, something that studies say is important for various reasons. Often, when stuck on plots or characters, I write stuff down in cursive. It’s different than typing and, in my case, very helpful in the creative process.
Bonus No. 1: A story written by ChatGPT
Write a short story about a man who’s virtually invisible to society.
John had always felt invisible. He was a quiet man, unremarkable in almost every way. He had no outstanding features or talents to speak of, no great achievements or accolades. He worked an unremarkable job at a small accounting firm, and when he went home at night, he spent his time alone, reading or watching TV.
It wasn’t until John lost his job that he truly began to feel invisible. He applied to dozens of new positions, but no one seemed to notice him. His emails went unanswered, his phone calls unreturned. It was as if he had ceased to exist altogether.
John tried to keep busy during the day, but his search for work had grown increasingly futile. He started to spend more and more time wandering the city, blending into the crowds of people on the street. No one ever looked at him or acknowledged his presence.
As the weeks went on, John’s invisibility became a sort of superpower. He could move through the city unnoticed, listening to conversations and observing people without fear of being detected. He saw the city in a different way than he ever had before, noticing the small details that most people ignored.
But despite this new power, John still felt alone. He longed for someone to see him, to recognize his existence. He started to leave little notes around the city, in the hopes that someone would stumble upon them and reach out to him.
It wasn’t until months later that John’s wish came true. He was sitting in a park, reading a book, when a woman approached him. She had found one of his notes and was moved by his story. They struck up a conversation, and John was surprised to find himself opening up to her in a way he never had before.
From that day on, John was no longer invisible. He had found someone who saw him, who recognized his worth. And as he walked through the city with her by his side, he realized that he had been invisible all along not because of any inherent flaw within himself, but because no one had taken the time to really see him.
The story has elements of Conagher, and it’s not too bad. A bit simplistic and a bit soppy, but then I don’t mind simple soppy stories . . . but it could use more guns, that’s for sure.
The writing reads a bit flat, partly because it has no ‘breaks’.
That’s something I consciously do in my writing, usually by adding humor. It may be subtle, or it may be just in passing, but humor is in almost everything I write.
I’m not convinced ChatGPT could infuse humor in a piece, at least not without very specific prompt directions. Heck, I’m not even sure it would recognize humor . . . or would it?
ChatGPT example 4
Tell me what makes this joke funny: A man goes to a zoo, but the only animal in the entire zoo is a dog. It was a Shih Tzu.
The joke is funny because the punchline is unexpected and plays on the audience’s expectations. The setup of the joke leads the audience to believe that the zoo will have a variety of animals, but the punchline reveals that the only animal in the zoo is a small dog, which is unexpected and amusing. Additionally, the use of the specific breed of dog, “Shih Tzu,” adds an extra layer of absurdity to the joke.
So, I’m guessing you don’t do puns?
As a language model, I can understand and generate puns, but my understanding of humor is based on patterns and statistics in the text I was trained on. Puns are a type of wordplay that rely on multiple meanings of a term or on a word sounding similar to another word. They can be a source of humor, but it also depends on the context and the audience.
The dog’s name sounds like an unflattering descriptor for the zoo.
Yes, that is another aspect of the joke. The word “Shih Tzu” sounds similar to “shi* zoo” which is an unflattering descriptor for the zoo. This adds another layer of humor to the joke, as it highlights the inadequacy of the zoo and the disappointment of the visitor. It also play on the audience’s expectation of a zoo having multiple and diverse animal exhibits.
Very good. How about this: Time flies like an arrow, but fruit flies like bananas.
This is a pun that uses the multiple meanings of the phrase “flies like.” The first half of the sentence “Time flies like an arrow” is a metaphor that means time passes quickly, like an arrow in flight. The second half “but fruit flies like bananas” is a pun because “flies” can also refer to the small insects commonly found near ripe or rotting fruit, and it is saying that these insects are attracted to bananas specifically. This creates a humorous contrast between the metaphorical meaning of the first half of the sentence and the literal meaning of the second half.
A bit clinical but not too bad.
Conclusion
So, it’s established I wouldn’t use ChatGPT and that I’m skeptical of its practical use beyond a narrow set of circumstances and not for the long term.
I’m sure many will adopt it, especially as it improves. By the way, the free version is getting restricted, but you can pay $20/month for unlimited use.
Will readers be able to tell when someone adapts ChatGPT or an equivalent AI as part of their workflow?
My guess is that the majority of readers will not be able to spot the difference and will likely enjoy the resulting content without caring how it was originated. It will be up to the writers to come clean regarding their use of the ‘tool.’
So, why did I say I was mildly troubled about ChatGPT and other upcoming AIs?
Because I think we will lose something that’s hard to describe.
For one thing, I don’t think AIs are innovative, at least not yet. One can argue it doesn’t matter because all stories pertaining to the human condition have already been told multiple thousands of times. Still, stories evolve along with the times they inhabit, told in ways that reflect the then-current state of society and its members.
That requires recognition of the spectrum of the human condition at a particular point in time, and I don’t think AI can develop such recognition. Worse, they might only focus on select portions of said human condition as recognized by their programmers.
I’m as suspicious of entities lacking empathy just as I’m suspicious of the many sociopaths in positions of power making decisions affecting millions without any indication they feel any guilt about the harm they cause others.
Plus, if AI algorithms have proven anything in the last decade is that they can capture and affect the attention of humans to negative consequences. They don’t feel anything, so they’re not malicious per se . . . it’s just how they are programmed. And their programming isn’t to grow society and improve the human condition; it’s to maximize profit.
Sure, there are always initial lofty goals, but history shows they always fall to power and greed (I’m looking at you, Google). Some good remains, but there’s also harm, and which is greater isn’t clear until much later.
I don’t expect this innovation to be any different.
That’s it. This post has ended . . . except for the stuff below.
<><><><o><><><><><o><><><>
Note: if you are not reading this blog post at DisperserTracks.com, know that it’s copied without permission, and likely is being used by someone with nefarious intentions, like attracting you to a malware-infested website. Could be they also torture small mammals.
Note 2: it’s perfectly OK to share a link that points back here.
<><><><o><><><><><o><><><>
If you’re new to this blog, it might be a good idea to read the FAQ page. If you’re considering subscribing to this blog, it’s definitely a good idea to read both the About page and the FAQ page.
Disperser Tracks 
I’m reminded of an early ’70s movie “Colossus: The Forbin Project” in which computers take over the world. It’s a little dated now but it seems more relevant now than in the ’70s. I think this AI technology is very interesting
Btw: I will be switching to a new PC (probably over the weekend). It may take considerable time to get running but I will try to keep up with my Chromebook.
LikeLiked by 2 people
That was a conscious entity with, if I remember correctly, very dictatorial tendencies (not unlike humans).
Really, I think the message was that even if we get to have robot overlords, it will make little difference to the lives of ordinary people.
As for keeping up, don’t worry about it, and good luck with the new PC.
LikeLike
I’m left wondering what the difference is between politicians who employs staff to write speeches and letters on their behalf and you or I using AI to write a blog page draft for us. I can’t see too much harm in it but nor would I bother to to do it.
Nevertheless, it’s interesting to see how AI responds to the traps you set for it.
LikeLike
There is no difference . . . in both cases, it’s aimed at presenting oneself as someone different than one’s actual self.
Really, it’s what seems to drive most of the Internet’s popular sites.
I suppose I do the same thing; I present myself as an opinionated and grumpy old guy who writes both fiction and non-fiction pieces and posts them along with lengthy seldom-viewed posts featuring some of the thousands of photos I snap every year.
Oh, wait . . .
LikeLiked by 1 person
I think it’s called “delegating” or “out-sourcing”.
LikeLiked by 1 person
Too bad Carnevale is over . . . I would have liked to paint my face like the last photo.
LikeLike
Hmm . . . you have enough issues with your eyes without painting your eyeballs. Even that much paint around the eyes is a tad risky.
LikeLiked by 1 person
No worry . . . I’d have to have some major change in my thinking in order to even contemplate doing that!
LikeLiked by 1 person