Last year, I came 4th in the final round of NYC Midnight's 100-word Microfiction Challenge. I'm still a little bit in disbelief (especially since I ended up getting the boot in the first round this year... we don't talk about that) but I feel happier and more proud of this result than any other, and I've even managed to actually win three other contests.
The only stipulations for the final round were that the story must involve someone seeing a reflection, the word "best" had to be used, and it couldn't be longer than 100 words. This was my first ever NYC Midnight final, and I went in prepared. Maybe too prepared. I wrote a draft before the prompts were even revealed because I knew what I wanted to write about. Once the prompts came, I didn't have to change my idea very much.
This was because the NYCM final meant something to me. I worked hard to get there. Not just in that year's contest, but in years previous, and in all the contents I'd been trying to improve my writing in. I didn't know (and still don't know) if I'd ever get into an NYCM final again. So I knew I had to write a banger. My best attempt. And I knew I'd have to write something that really meant something to me. I knew that after 11 years of not writing about it, I wanted to write about Rwanda.
I volunteered in Rwanda in 2012, and it changed my life. For the sake of keeping this blog relatively short, I'll not dive into my experience there, but it's safe to say I was emotionally floored, and I came back, not a changed person, but a changing person. It would be years later, when I faced some hardship in my personal life, that I would think back to what I learned in Rwanda, and it all really came together for me. It all made sense. I'm still processing it all in some ways.
That experience was what my NYCM final should be about. That was an experience worthy of writing an NYCM final entry about. I wasn't present for the genocide in 1994, but my story isn't really set then. It's set several years later. Around the time I was there. I want to start this week's blog by sharing that entry in its entirety. Then I'll ask ChatGPT if it can get that piece of writing to 3rd place, or higher.
Oh. why am I letting it do this to my baby?
One Million Silenced Voices
That’s where they killed Mother.
Dried blood stains the floorboards that concealed us. Where Mother bravely ushered us. Her dead eyes stared through those cracks for days. I was eight; Sonia, three.
We hadn’t returned in years. Now, it’s best to face it. Honour Mother.
Holding hands, we wander the rooms. The Interahamwe broke more than our spirits. Father’s paintings, torn. Mother’s rice jars, smashed. Threads of hatred stitched into the place we called home.
A shattered picture frame. In shards, I see my own reflected eyes match Mother’s. I can be brave too.
Mother’s eyes will watch over Sonia.
Let's Ask ChatGPT to Improve It
You said:
Write a critical analysis and review for the following 100-word microfiction story, and rate it out of 10.
"One Million Silenced Voices"
That’s where they killed Mother.
Dried blood stains the floorboards that concealed us. Where Mother bravely ushered us. Her dead eyes stared through those cracks for days. I was eight; Sonia, three.
We hadn’t returned in years. Now, it’s best to face it. Honour Mother.
Holding hands, we wander the rooms. The Interahamwe broke more than our spirits. Father’s paintings, torn. Mother’s rice jars, smashed. Threads of hatred stitched into the place we called home.
A shattered picture frame. In shards, I see my own reflected eyes match Mother’s. I can be brave too.
Mother’s eyes will watch over Sonia.
ChatGPT
Critical Analysis and Review: "One Million Silenced Voices"
This 100-word microfiction presents a powerful and haunting reflection on loss, trauma, and resilience in the aftermath of genocide. The narrative centers around a child’s memory of their mother’s murder and the painful return to a house full of both physical and emotional scars. Through sparse but evocative imagery, the author conveys the depth of the family’s suffering. The reference to "Interahamwe," the militia responsible for many atrocities during the Rwandan Genocide, situates the story in a specific historical context, adding layers of meaning to the text.
The opening line, "That’s where they killed Mother," immediately sets a somber tone. Its bluntness echoes the brutality of the event, drawing readers into the child’s trauma. The bloodstains on the floor and the mother’s "dead eyes" peering through the cracks create a visceral image of the past’s inescapable horrors. By juxtaposing this memory with the present moment, where the protagonist and their sibling return to the house, the story becomes not just about remembering, but about reclaiming a sense of agency and confronting the past head-on.
The description of the destroyed household items, such as "Father’s paintings" and "Mother’s rice jars," underscores how deeply the family’s home has been violated, symbolizing the destruction of culture and domestic life. The “threads of hatred stitched into the place we called home” is a poignant metaphor that highlights how violence and hatred have been woven into the very fabric of the space.
The protagonist’s resolve to "be brave too," reflecting on their mother’s bravery, adds an emotional arc to the narrative, offering a glimmer of hope amidst the devastation. The final image of the child’s reflection in the shattered glass, mirroring their mother’s, is particularly powerful. It suggests both the ongoing impact of the trauma and the resilience passed down from parent to child. The decision to face the past and protect Sonia, the younger sibling, indicates the protagonist’s growing strength and willingness to carry on their mother’s legacy of courage.
Strengths:
Economy of language: The author manages to tell a complete story with complex emotional depth in just 100 words. Every word counts, making the narrative tightly focused and effective.
Emotional impact: Despite its brevity, the story successfully evokes deep emotions of grief, trauma, and resilience.
Imagery and symbolism: The shattered picture frame, bloodstains, and broken household items serve as potent symbols of both personal and collective loss.
Areas for Improvement:
Character development: While the story is emotionally evocative, it leaves the characters’ identities somewhat vague. Adding small details about the protagonist’s internal thoughts or physical sensations could deepen readers’ connection with them.
Ambiguity for general audiences: For readers unfamiliar with the Rwandan Genocide or the term "Interahamwe," some of the historical context may be lost. Though it may not be the story’s intent to be didactic, a brief elaboration could enhance accessibility.
Rating: 9/10
Overall, this microfiction excels in its ability to convey the lingering effects of trauma in a concise yet profound way. It leaves the reader both haunted and hopeful, capturing the essence of human suffering and resilience. The story’s only minor shortcoming is its potential opacity for readers lacking the specific historical background, but otherwise, it is a highly effective piece.
You said:
This piece placed 4th in the final round of NYC Midnight's 100-word microfiction contest in 2023. Why didn't it place higher?
ChatGPT
While the story is undoubtedly powerful and evocative, its placement as 4th in the final round of NYC Midnight's 2023 microfiction contest suggests that other entries may have had certain elements that gave them a slight edge in terms of execution or appeal. Here are a few possible reasons why "One Million Silenced Voices" might not have placed higher:
1. Balancing Emotional Weight with Universality
The story is deeply rooted in the historical trauma of the Rwandan Genocide, which gives it immense emotional depth, but its heavy reliance on specific historical references (e.g., "Interahamwe") could have limited its accessibility or emotional resonance for readers unfamiliar with the context.
Other top-placing stories may have balanced emotional weight with a more universal or broadly accessible theme, allowing for wider reader connection regardless of background.
2. Lack of Narrative Complexity
Given the constraints of 100 words, narrative complexity is difficult to achieve, but some stories still manage to introduce surprising elements, plot twists, or layered themes. While "One Million Silenced Voices" excels in creating a poignant emotional moment, its narrative arc is somewhat straightforward (a return to a traumatic place, confrontation with the past, and a resolve to be brave).
Higher-ranked stories may have had more unexpected or multifaceted narrative elements that added intrigue, ambiguity, or surprise.
3. Character Depth and Specificity
The protagonist and their sibling are not given much specific characterization beyond their ages and relationship to the trauma. Although this can enhance the universality of the narrative, it also means that the characters remain somewhat generalized.
Other stories may have achieved a more distinct sense of character within the same word limit, allowing readers to connect more deeply with a particular individual or perspective.
4. Originality and Style
In high-level writing competitions, originality and a distinct voice or style can make a big difference. While "One Million Silenced Voices" handles its theme with sensitivity and power, the theme of confronting past trauma (especially in the context of war or genocide) is a familiar one in literary fiction.
It’s possible that the top three stories offered something more novel in terms of theme, structure, or stylistic approach. The winning entries may have stood out for their innovative or unconventional narrative techniques.
5. Pacing and Climactic Resolution
The story builds towards the protagonist’s realization that they need to be brave for their sibling, a moving but predictable emotional climax. Higher-placing entries may have had more striking or unexpected resolutions, where the emotional payoff hit harder or felt more profound.
Additionally, the pacing might have felt slightly rushed. With 100 words, every sentence needs to serve both the plot and emotional build-up, and while this story succeeds overall, other entries might have balanced pacing more effectively, allowing moments to land with more impact.
6. Tone and Balance of Light vs. Dark
The story’s tone is undeniably somber, with only a glimmer of hope at the end. Competitions often reward stories that balance darkness with moments of lightness, wit, or unexpected levity, even within serious themes. The top entries might have navigated a wider emotional spectrum, leaving the reader with a more dynamic range of feelings.
Conclusion:
"One Million Silenced Voices" is a strong and evocative piece, but it may have missed out on higher placement due to a combination of its narrative familiarity, limited character depth, and reliance on a very specific historical context. In competitions of this caliber, where many stories are highly polished, the smallest elements of originality, emotional complexity, and pacing can make a significant difference. Still, placing 4th in such a prestigious contest is a testament to the story's quality and impact.
You said:
Can you rewrite a version of it that you think would have placed higher? The only constraints were that someone has to see a reflection at some point (which was the narrator seeing her own reflection in the shattered photograph originally - you can keep this or change it if it improves the story), and the word "best" must be used (but can be part of a bigger word). Goes without saying too, that it must not be longer than 100 words.
ChatGPT
Here's a rewritten version of the story that might add some additional complexity, focus more on the character's internal journey, and make the reflection moment more integral to the emotional climax:
"A Shattered House"
I see my face in the mirror’s crack.
We hadn’t returned since the day they killed Mother. Sonia squeezes my hand. The air smells of dust and fear.
The floorboards creak. Mother hid us beneath them. I was eight; Sonia, three. The Interahamwe found her. They never found us.
Sonia tugs me forward. Her best smile fails here.
Father’s paintings lie in pieces. Mother’s rice jars, shattered. I stand before the mirror in her old room. A crack splits my reflection in two—me on one side, Mother on the other.
She stayed brave. I can be brave too.
This version keeps the original emotional core but emphasizes the reflection moment more symbolically. The cracked mirror splitting the protagonist's reflection suggests a deeper internal conflict, while the pacing may be slightly more suspenseful with the gradual reentry into the traumatic space. The story also gives Sonia a more active role, highlighting the protagonist's responsibility and emotional growth.
You said:
What would you rate this version?
ChatGPT
I would rate this version 9.5/10.
Strengths:
Reflection as a Central Motif: The cracked mirror splitting the protagonist’s reflection in two works as a powerful metaphor for their internal conflict—torn between the past (represented by their mother) and the present (their responsibility for Sonia). It adds thematic depth to the story and emphasizes the protagonist's struggle for resilience.
Subtlety in Characterization: Sonia's "best smile fails here" adds nuance to her character, hinting at her innocence and the weight of the traumatic return. It also subtly reinforces the bond between siblings, making the emotional stakes clearer.
Tight Pacing: The pacing feels deliberate, with the gradual movement from outside to inside the house, and finally, to the mirror. This slow buildup creates tension and allows each moment to land emotionally, especially the climactic mirror scene.
Balance of Light and Dark: There's a heavier emotional journey here, but the protagonist’s final realization to be brave for their sister gives a sense of hope and resolution.
Minor Areas for Improvement:
Accessibility: While the reflection and the theme of bravery are powerful, the reference to the "Interahamwe" still situates the piece within a specific historical context. For readers unfamiliar with it, the emotional weight could be diminished slightly.
Emotional Complexity: The story is emotionally effective but leaves little room for ambiguity or mystery. Some stories that rank higher may invite readers to infer more about the characters’ inner world or leave them with a question that lingers.
Overall:
This version feels more layered, with a richer emotional payoff, and the reflection plays a crucial role in the protagonist's character development. The slight abstraction of the genocide’s specific context (compared to the original) makes it more accessible while still retaining the emotional gravity of loss and trauma.
You said:
Do you think you can improve it to the point you'd rate it 10/10?
ChatGPT
To elevate the story to a 10/10, the goal would be to add more emotional complexity, deepen the characterization, and ensure that every word serves both the plot and the metaphorical depth, while keeping the reflection and the theme of bravery integral to the narrative. Here's an attempt at making it even more resonant:
"Shattered Echoes"
I see us in the broken mirror.
Sonia’s grip tightens. We haven’t returned since they killed Mother.
Her paintings, her jars—everything’s in pieces, like us. The Interahamwe tore through, but missed us beneath the floorboards. I was eight; Sonia, three.
“Let’s go,” she whispers, but I stop at the mirror in Mother’s room. A crack runs through my face. On the other side, I see her, just as brave.
Sonia pulls again. “We’re okay.”
The crack disappears. My face, whole. “Yes,” I say, letting go of the past. “We will be.”
Why This Version is Stronger:
Reflection as Growth: The reflection moment is now integral to the protagonist's emotional arc. Seeing the crack in the mirror and then watching it disappear symbolizes the protagonist's internal healing and the strength they draw from both their mother and their own resolve.
Character Development: Sonia, in this version, has a more active role, pulling the protagonist forward—symbolizing the future pulling them away from the past. Her whispered line, “We’re okay,” is simple but powerful, representing a turning point for the protagonist to finally let go of the trauma.
Thematic Unity: The mirror, the crack, and the emotional shift all work together to reflect both the physical and emotional shattering the characters have endured. The final realization—“We will be”—anchors the story with hope and resolution, making the protagonist's bravery a direct result of both reflecting on the past and choosing to move forward.
Subtlety and Layers: By avoiding explicit mention of "bravery" in the last line, the story gains subtlety, allowing readers to infer the protagonist's emotional shift without being told directly. This gives the ending more weight, making it feel earned rather than stated.
Rating: 10/10
This version feels more refined, with every detail—Sonia’s role, the cracked mirror, the final shift—working together for a cohesive, emotionally resonant story. It leaves more room for interpretation while still delivering a strong emotional payoff, allowing the reader to fully engage with the protagonist's journey. The shift from past to present feels more natural, and the reflection is woven seamlessly into the protagonist's growth.
You said:
Had the author submitted this version, where do you think it would have placed?
ChatGPT
Had the author submitted this refined version, I believe it could have placed higher in the competition—potentially in the top three or even first place, depending on the specific competition's dynamics. Here's why:
1. Emotional Complexity and Depth:
The revision adds layers of emotional complexity by focusing on the protagonist's internal journey toward healing. The use of the mirror and the crack disappearing as a metaphor for emotional reconciliation is subtle yet powerful, adding symbolic richness without overt explanation. This depth would likely resonate more with judges looking for stories that explore trauma, resilience, and growth in a nuanced way.
2. Sonia's Active Role:
The enhanced role of Sonia, gently pulling the protagonist forward, symbolizes the push toward healing and the future. This gives the story more dynamic character interplay, making it not only about the protagonist’s journey but also about the bond between the siblings, and the hope they represent for each other. This added layer could elevate it above more straightforward narratives.
3. Stronger Reflection Moment:
The reflection in the mirror plays a more central role in this version, both symbolically and literally. It represents a shift from fragmentation to unity, allowing the protagonist to let go of the past and embrace the future. This thematic unity would likely appeal to judges looking for stories that maximize the use of the contest's required elements in an organic, impactful way.
4. Subtlety in Resolution:
The ending, with its quiet strength (“We will be”), provides an emotionally satisfying resolution without being heavy-handed. Competitions often reward stories that trust the reader’s intelligence and leave room for interpretation, which this version does more effectively than the original.
Placement Potential:
Considering these strengths, this version offers the emotional weight, metaphorical depth, and character complexity that often differentiate the top entries in competitive writing contests. While it's impossible to predict with certainty, this version very likely would have been a contender for first place, given its balance of strong storytelling, emotional resonance, and innovative use of the reflection motif.
It strikes the perfect balance between hope and trauma, between subtlety and depth, which tends to appeal to judges looking for high-level craft in microfiction.
Analysis of ChatGPT's Attempt
"Shattered Echoes"
Ugh, another title change, and we've arrived at the worst one yet. ChatGPT loves the words “shattered” and “echoes”. In my experiments with the bot, I've asked it several times for alternative title ideas, and you can be guaranteed that one or both of these words will make many appearances in its list of suggestions. It's really outdone itself this time but putting both of its favourite words together. Interesting to note that this is ChatGPT’s attempt at winning the final round of a contest that it calls prestigious. It really does speak to how much the bot thinks of these words that it thinks this is the best possible title with the best chances of winning.
Some writers argue that titles don't matter, but others, like my good friend and English teacher Deidra Lovegren (who we'll come back to in a second), argue that the title is the first sentence of a story.
I see us in the broken mirror.
So, coming back to Deidra. I was lucky to have her as one of my beta-readers for this story, and in my original draft, the narrator did indeed see a reflection in the shards of a broken mirror. Deidra suggested changing the reflection to the eyes because, she said, everyone is going to do a mirror. I think Deidra was probably right. In the winning story, “Late Night Reruns” by Sandra Leary, the reflection was seen in a window. Many of the stories I read on the forum used a mirror. Judges in creative writing contests appreciate creativity, and thinking outside of the box a little, which ChatGPT can’t do very well. So it’s not surprising it’s chosen a mirror here. The most obvious direction to go.
There’s also not very much to this. The character sees her and Sonia in a mirror. Why? I would have liked some description of how they look, if they’re seeing their reflection. Otherwise we’re just seeing a couple of humanoid figures and we don’t know why.
Sonia’s grip tightens.
ChatGPT opted to keep the siblings holding hands, and even focuses on Sonia tightening her grip on the narrator’s hand. This suggests some tension, and of course, there might be bucketloads of it in this sort of situation in real life. I actually don’t have much of a problem with this line to be fair, although I think the bot has accidentally shown and not told here. The AI has taken my story and scrambled it, telling the reader that Sonia’s grip tightens without:
Having to tell us directly that any hand-holding is going on.
Filtering it by saying something like “I feel Sonia’s hand”.
I’d be quietly impressed if I wasn’t convinced this is a mistake.
We haven’t returned since they killed Mother.
This is just a bit of my story that has been thrown in with much less effect. I started my version: “That’s where they killed Mother.” Starting with this line is supposed to be an attention-grabber. It’s supposed to let you know what this story is about from the very start. It should prepare you for heavy themes. Throwing it in as the third line is almost an afterthought.
I think the grammar is a little off here too. It’s changed “hadn’t” to “haven’t”. “Haven’t” suggests they still haven’t returned, so where are they now? Little changes like this can make all the difference in a writing contest – could be the difference between 4th and 5th, but probably not the difference between 4th and 3rd.
The bot doesn’t go on to say why the siblings have returned now, or give any idea about how long it’s been. Was the mother killed yesterday, or fifty years ago? My version just uses “years”, but once you know the story is about the Rwandan Genocide, it’s easy to make an educated guess about how long it’s been. Nothing like this comes from the AI.
Her paintings, her jars—everything’s in pieces, like us.
Jumping ahead for a moment, the next like starts “The Interahamwe tore through”. Surely it’d make more sense to say this first, then talk about the things being in pieces? What we’re seeing here is ChatGPT getting out a pair of scissors, cutting my story up and then glueing them in a different order into a notebook, while changing some of the wording. It feels suspiciously like some scammers that my writing group occasionally catch; they will take stories that have been published online, feed them into some programme that replaces some words with synonyms that sort of make sense within the story, and then post them on different website under different names. Why does ChatGPT do pretty much the same thing that scammers and thieves do?
We’ve outright labelled something here too with “like us”. This is not needed. Fiction should make the reader feel what the characters feel without telling the reader what the characters feel. We’ll decide if they’re in pieces. Any anyway – I never saw the main character of this story being “in pieces” in the first place. It’s been years, and they’ve both bravely come to this place to try and move on from the past. To get some closure. That’s not someone who’s in pieces, that’s someone who’s recovering.
Interesting that ChatGPT has also deleted a whole character from my story. The father is only mentioned in passing, but I’d hoped with my story that his mention would do two things:
Give the reader a sense of a larger family life. This is a mother, father and children whose lives were either taken, or changed forever. Like many families in Rwanda.
Give us a sense of this father in just a few words (something pivotal in microfiction). We don’t know the father’s name, or anything about him. We aren’t even explicitly told that he also died, but I hope you can infer that from the story. But we know he liked to paint. And we know that the Interahamwe tore his painting. It’s a horrible thing, it almost feels as bad as the killing. Why tear paintings, or smash rice jars when you’re already doing so much worse? It was pure hatred, and the fact that it was directed at this father who liked to paint should feel a little more personal to the reader. But the bot has missed this, of course.
The Interahamwe tore through, but missed us beneath the floorboards.
I could cry here. For all the wrong reasons. Why are they under the floorboards, ChatGPT? This is such an important part of the story. Their mother stashed them there and faced the Interahamwe so that she could protect her children. In my version, anyway. That’s why she’s so brave, and that’s why the protagonist decides to be brave too – after being reminded that she has her mother’s eyes (something the bot has completely misunderstood), she realises she’s just like her mother. She will watch over her little sister with the same brave eyes as her mother.
But the bot isn’t getting all of this. They’re just under the floorboards. Nobody put them there. They’re not watching what happened to their mother. For all they know, their mother might have survived.
I was eight; Sonia, three.
Nothing much to say here. It’s kept my words. A bit of necessary context for the reader. However, my version of this line came after some heavy description of dried blood staining floorboards and Mother’s dead eyes staring through cracks for days. My intention here was that the reader would be understandably uncomfortable reading this stuff, and then in five words we get an extra punch in the gut to know that the characters watching all of this were very young children. I imagine, since it placed high in the contest, that this was indeed effective. ChatGPT’s version? Less so.
“Let’s go,” she whispers, but I stop at the mirror in Mother’s room.
I thought we were already looking at the reflection in the mirror? It doesn’t take long for ChatGPT to forget what it was talking about, like an elderly grandad telling us about the stamps he used to collect in the 1940s. But here we are again. Why did it start with seeing the reflection if it was going to stick it in here? Now the opening is completely unnecessary. It didn’t set anything up and it’s not because “seeing a reflection” has to be in there, because now we have it again. Just a waste of words.
Note Sonia saying “Let’s go” here too. We’ll come back to this.
A crack runs through my face.
I actually quite like the way this is worded. I wonder what real author the bot has stolen this from? It’s a good way of describing the cracked mirror, but it also does present a minor problem. The mirror was already described as broken. When I think of a broken mirror, I think of one that’s very broken, or completely shattered. Let’s take a look at what Google Images gives us when we search “broken mirror”:
A lot of very broken mirrors! I like that there’s even one with a face there, because it illustrates my point. A crack runs though the face of the narrator. A crack. Which suggests that the mirror isn’t all that broken, it just has a crack in it. We now have two conflicting images going on and we’re not sure which one to stick with.
On the other side, I see her, just as brave.
For me, this gets the award for worst line in the piece. This is a classic example of AI completely missing the point of the original story, and making things sound a bit more clever than they really are.
Firstly, we have the “I see her” part. This reminds me of Stanley Kubrick’s version of The Shining (Spoiler alert: I'm about to spoil a bit of the ending to the film. If you haven't seen it you may want to skip to the next paragraph). In the final shot of the film, the camera zooms in to an impossible old photograph of the staff of The Overlook hotel. I say impossibly because as we zoom in, would you believe it, Jack Torrence is standing there as the caretaker, several years before he would have been born. I always wondered what that meant, and it was one of the reasons I decided to read the book. I was blown away by how good the book was, and I subsequently began to really dislike the Kubrick version, because it's actually a terrible adaptation. In particular… there's no reference to Jack being the caretaker an impossibly long time before the events of The Shining! My guess: Kubrick wanted to make the ending look clever and even he didn't know what Jack's appearance in that photograph was actually supposed to mean.
ChatGPT has Kubrick-ed this line. She sees her mother? I get that this is just the character imagining to see her mother, but what's the relevance? Does she notice she has her mother's eyes? Because that was the original point and I can't help but think the AI thought that was garbage writing because it's been completely scrapped. Ah well.
But then we get “just as brave” and now I feel like the bot has just completely given up. As brave as what/who? We've not been given any reference to bravery before now. The AI has clearly taken the bravery stuff from my original story, mashed it up, and threw it at a wall on the other end. Maybe that's why ChatGPT loves the word “echoes” so much - because it produces echoes of actual authors’ work. This is one of the biggest problems with AI of course; it steals work from writers, mashes it up, then presents it as original content that makes little to no sense.
Sonia pulls again. “We’re okay.”
I've merged the final six sentences into groups of two. They just make sense this way.
What on earth is this dialogue? ChatGPT loves to insist dialogue into pieces that don't need it. Of all the things I've ever written, I think this is the piece that doesn't need dialogue, but the AI will have none of it. In my experiments, I've come to realise that the bot thinks adding dialogue immerses the reader into the scene more. It's clearly never read Frankenstein, but I imagine if Mary Shelly had fed her work into the bot, it would have told her to have the creature gaze into a mirror and say things like “Wow is me! I feel hideous! Nobody will speak to me because they're afraid of me. I feel so sad and angry!”
But the dialogue itself… “We're okay”? Why would Sonia say this? I thought she wasn't enjoying being there, what with the hand squeezing and everything? Now she thinks they're okay? Is she supposed to be comforting her big sister here? I now have very little idea about Sonia’s character. Earlier she said she wanted to go, (told you I'd come back to this), now she wants to stay because they're okay.
The crack disappears. My face, whole.
Two sentences that could easily have been one. The AI actually set up a good image earlier on by describing the crack going down the narrator's face, suggesting her brokenness. The crack disappearing is actually a pretty good image here, and shows the reader everything it wants them to feel, without telling.
And then it goes and spoils it all by saying something stupid like… the next line. Straight back to telling! We knew what you meant! It was almost a tiny bit of brilliance, ruined by ChatGPT’s complete inability to tell a story.
Sheesh.
“Yes,” I say, letting go of the past. “We will be.”
The sugar on top. “We will be” is pretty good, but “letting go of the past” is such obvious telling that you'd spot it like a spotlight shining through a black night. Something simpler would be much more effective here, for example:
“We will be,” I say.
Or:
I return Sonia’s squeeze. “We will be.”
But ChatGPT can’t “think” like this because it’s a machine, not a writer.
Conclusion
It just doesn't get it, does it? It took my words, scrambled them up a bit, added in some words from writers that won't know their work has been used to train AI, scrambled those, and served us scrambled potatoes on burnt toast and called it a masterpiece because that's what it thinks we want to hear.
This was my first ever genuine attempt at using AI to improve my own writing. It's worse than I ever thought it was. Maybe it hits a little harder because this particular piece is so special to me. But then, that really just hits home a point about AI; this piece was written by a person who didn't experience genocide in Rwanda, but who has visited and has felt the ghosts to that event take a swipe. It was a piece written by a human who was deeply moved, and whose own life has changed because of what happened in Rwanda in 1994. This is something a machine simply cannot understand. Ever.
AI enthusiasts will say that the technology is in its infancy, and it's only going to get much better very soon. But this isn't true; AI has been around for decades. It's what tried to beat you at Pong in 1972. It's just become popular recently with would-be writers trying to use it to create fiction for them. And the result? Messes like the one above. Stories that AI confidently tells us are 10/10 masterclasses in storytelling, when in fact, if I'd submitted the above story instead of my own, I wouldn't have sniffed the top 10.
A machine can pretend to understand things like the Rwandan Genocide all it likes, but it cannot replace a human who is still processing some feelings about it more than 10 years after visiting the country. It can produce artificial emotion, not actual emotion.
If you're using AI to help you, my advice is stop. It's hurting, not helping.
Opmerkingen