I recently did an experiment where I went head-to-head with ChatGPT, writing a science fiction novel to compare and contrast an A.I.’s approach to fiction as opposed to a human author. (My sci-fi book is here.)
The Prompt:
Write a science fiction novel opening from a first-person point of view.
That was the only prompt I gave ChatGPT and the only direction I took myself. Here’s what we each came up with.
ChatGPT’s Opening:
“I woke up with a jolt, disoriented and surrounded by a cold metallic hum. The last thing I remembered was the blinding flash of light and the deafening roar of an explosion. As my vision cleared, I realized I was lying in a sleek, futuristic chamber filled with pulsating screens and unfamiliar technology.”
My Opening:
“You don’t remember taking your first breath, but you sure as hell remember taking your last. The rattle of pus in your lungs. The explosion in the base of your skull as nanites cross the blood-brain barrier and begin devouring the soft flesh of your fragile brain. They say morphine disrupts the transfer of consciousness in the moments after death, and the machines need us to feel every bit of it for the damn thing to work. I call BS. Tech enjoys watching humans suffer like the lab rats we are… or were in my case.”
My Analysis of ChatGPT’s Opening:
1. Chat GPT begins with the character waking up
In my view, ChatGPT makes a common writing mistake by having the character wake up in the first paragraph of the novel. I have written a previous article about this rule (and also made a video about the ways to break it.) While it could have ultimately worked in this particular instance because the character is waking up to something exciting happening (ie. being trapped in a futuristic chamber), the problem is the lack of emotion.
2. There’s no emotion, stakes, or urgency
If a human character woke up with their body jammed inside a tube and no memory of how they got there, they might feel claustrophobic, panicked, afraid. They may be sweating, kicking, or clawing for a latch to try to get themselves out. Chat GPT’s character’s reaction is completely sterile and devoid of any authentic emotion. Instead, we as the reader are just left with a vague observation of what the tube is like rather than being plunged into the experience of being stuck inside a tube.
3. No sense of voice
My second impression of the opening is that we don’t know anything about the character in the tube. We don’t have a sense of their voice.
When Katniss Everdeen woke up at the beginning of The Hunger Games, we knew in the first paragraph that she has a sister that she’s protective of. We immediately get a sense of who Katniss is. I’m not saying that a character’s entire personality has to be revealed in the opening lines, but we should get at least a hint of who they are.
4. Filter words and vague descriptions
These prose seem unpolished to me.
For starters, Chat GPT used a filter word “realized” in the sentence “As my vision cleared, I realized I was lying in a sleek, futuristic chamber.”
Filter words are something I work to limit in my own writing as they distance the reader from the character’s experience.
Also, describing their surroundings as “a cold metallic hum” and “sleek futuristic chamber” with “unfamiliar technology” felt vague and uninspired to me. What about the chamber is futuristic? Are there no seams in the metal? Or is the surface reflective? What does it look like? What does it feel like? Labelling something futuristic or unfamiliar doesn’t appeal to our senses or paint a picture of it in our minds.
Chat GPT’s Analysis of My Opening:
This opening to [Candice’s] science fiction novel is compelling and immediately establishes a sense of intrigue and urgency. The vivid description of the protagonist’s last moments and the technology involved creates a visceral and unsettling atmosphere. The use of nanites and their role in the transfer of consciousness adds a futuristic and dystopian element to the narrative.
The narrator’s voice is strong and assertive, creating a sense of defiance or resentment towards the technology that manipulates human experiences. The mention of morphine introduces a layer of skepticism, suggesting a character who questions the motives and ethics behind the technology’s implementation.
The reference to humans as “lab rats” implies a broader theme of experimentation or manipulation by those in control, adding a potential layer of social or political commentary. The combination of technological advancements and the dehumanizing treatment of individuals sets the stage for a thought-provoking exploration of the relationship between humanity and technology.
Overall, the opening sets up a mysterious and dark world, leaving readers with questions about the nature of this advanced technology, the motives behind it, and the protagonist’s role within this dystopian landscape.
My Overall Assessment:
Chat GPT did a great job of analyzing my opening and picking up on themes while also identifying the voice and tone correctly. In my opinion, it did a better job of analyzing a piece of writing than it did coming up with original content, at least in this instance.
I’ve spent years cultivating my voice and writing style, and I know I’ll continue to evolve from here. But the same can be said for A.I.
Chat GPT’s opening looks a lot like the first paragraph of my first attempt at a novel, which makes me suspect that the A.I. can evolve (as I did) to do a much better job in the future. Time will tell, but I suspect with the entire compendium of human knowledge at its fingertips, it’ll evolve much faster than I did. It’ll be interesting to do another experiment like this a year from now to see how much it has changed.
The Impact of this Experiment:
I had such a great time writing this opening, that I am actually writing a novella around this concept (with this same opening I used here)! It is an adult sci-fi thriller that explores the relationship between humans and technology.
I’m currently writing the second chapter. If you would like to stay apprised of this new book, please sign up for my substack (it’s free) at https://candicejarrett.substack.com.