I Asked ChatGPT to Give Me Feedback on My Short Story — Here's How It Went

Subscribe to HubSpot's Next in AI Newsletter
Jay Fuchs
Jay Fuchs

Published:

I like to think I'm a good writer — I mean, I really hope so — but like virtually every other writer out there, I'm self-conscious, stubborn, and constantly questioning whether anything I put together is actually compelling or interesting.

chatgpt offering constructive feedback about a story about werewolves

It's hard to know for sure — and getting honest feedback to either validate my greatness, confirm my trash-ness, or affirm my okay-ness can be a struggle.

Friends can be reluctant to tell you that you suck. Peer feedback can sometimes be weirdly competitive, and the vast majority of people don't have the time or interest to sit down and get through a story of yours. So where do you go?

I‘ll be real. I held out on AI as long as I could. I was convinced it was going to be humanity’s downfall from the first time my phone had the audacity to tell me I actually meant to say, “I have no ducking idea” — but desperate times call for desperate measures.

I started a short story and wanted some quick, empathetic, valuable feedback — so I figured I‘d see what ChatGPT had to say. Here’s how that went.

Click Here to Subscribe to HubSpot's AI Newsletter

Table of Contents

The Story Itself

So for context, this is what I passed along to ChatGPT. It's a story about a world where werewolves exist and the government contracts private firms to exterminate them — in this reality, werewolf killing is a divisive political issue because half of the population sees being a werewolf as a medical condition.

If you‘re not interested in reading the story itself, I won’t take it personally — you can click here to see how ChatGPT approached it.

Here we go:

“A Force. A Friendship. A Family.”

Six words written in bold black lettering on the gold plaque positioned above the Silver Bullet Enterprises central Massachusetts branch's entryway — a sign that never made complete sense to Ronnie.

For one, the fact that a plaque for Silver Bullet Enterprises wasn't silver felt like a missed opportunity. Also, he figured a “friendship” generally exists between a handful of people at most — the fact that it was in reference to an entire office felt like they were scrambling to get an extra “F” in.

The only aspect of the sign that felt appropriate was “A Family.” The office was exactly like one of those to Ronnie — in that it was dysfunctional, embarrassing, and he didn't get to choose who else was a part of it.

It was a tough gig — werewolf extermination.

It was dangerous. It was mostly thankless. And it was the subject of great debate, vitriol, controversy, legislative battles, and litigation.

But on a near-daily basis, you got to kill something big without legal recourse. That was a big plus for a lot of people in the industry.

You‘d think a werewolf extermination service’s office would have some kind of mystique or campy magic to it — rife with cauldrons and spell books and witches with white eyes. Not the case at all.

Silver Bullet's central Massachusetts office was virtually indistinguishable from a headquarters for some midsize cybersecurity consulting firm: slate gray cubicles, the scent of stale coffee hanging in the air, creaky black desk chairs — all the hits.

Even though it dealt with the supernatural, Silver Bullet Enterprises was very much a business.

Every last bullet had to be accounted for. A dedicated legal team had to stave off the waterfall of lawsuits the company incurred. And the HR department had to coordinate sensitivity trainings that conveniently ignored the fact that everyone in the office worked for a business dedicated to murdering people — kind of. Not in a legal sense at least.

See once a person wolved-out, they were no longer legally considered a human being — despite the fact that they looked, talked, walked, and behaved like fully lucid, sentient people until the sun set.

Once it got dark, they forfeited control of their earthly form, grew to be somewhere between six and a half and seven feet tall, became covered in fur, developed razor sharp canines, grew claws on paws where their hands used to be, howled at the moon, and went into a murderous rage — typically killing livestock and pets.

For the most part, they didn't kill human beings — if you were to encounter one, it would probably just bite you and, in turn, make you prone to wolving-out yourself.

But that “for the most part” represented a pretty massive caveat — werewolves didn't kill adults, but they would rip a child to shreds without hesitation.

That's why werewolf extermination services like Silver Bullet had to exist — but not everyone saw it that way.

There was no cure for being a werewolf, but about half the population still saw it as a medical condition — and felt it should be treated as such.

People who became werewolves never completely lost touch with their humanity. In fact, they were human beings for most hours of the day, so by that side's account, werewolf extermination was flat-out murder.

That's why Ronnie passed by an ever-present horde of protesters on his way into the office every day — generally hurling names, insults, and the occasional object his way. He could reliably expect to take a few snowballs to the face over the winter and had narrowly avoided more than a couple of glass bottles during his tenure.

Ronnie was a little smaller than you'd expect an ex-marine to be — something he got plenty of crap for in basic training.

He stood about 5'7 with a slender frame, propping up a face with a sturdy jawline that he had gotten two whole compliments on over the years — ones that he ran back in his head nearly every day.

The same went for his “kind eyes” that the pretty bartender at CB Hannigans had pointed out. He tipped her a twenty that night — he didn't care that she was being paid to pretend to like him.

Today, he only passed two protesters as he pulled up to the office and was mostly able to tune out the particularly creative string of obscenities they threw his way. He‘d grown accustomed to handling names like that, and frankly, those ones were pretty tame compared to what he’d heard in the past.

He walked through the office's double doors, trading a gray, grody Worcester, Massachusetts February day for a workplace that was only slightly less gray and grody itself.

ChatGPT's Feedback

It took roughly 30 seconds for ChatGPT to produce all of this. Here's what it had to say:

Your story has a strong and engaging opening, and I appreciate the mix of humor and the darker aspects of the narrative. Here are some detailed points of feedback:

Strengths:

  1. Unique Concept: The idea of a werewolf extermination business run like a mundane corporate office is both unique and intriguing. It sets up an interesting juxtaposition that draws the reader in.
  2. Voice and Tone: Ronnie’s cynical and slightly humorous voice is well-established early on. This makes him a relatable and interesting protagonist.
  3. World-Building: You do a good job of quickly establishing the world and its rules, especially the legal and societal views on werewolves. This adds depth to the setting without overwhelming the reader with information.
  4. Imagery and Detail: Descriptive details like the office environment, the plaque, and the protesters help paint a vivid picture of Ronnie’s world.

Do I agree with ChatGPT?

I‘ll take literally any compliment anyone gives me about my writing (including robots), so take what I say here with a grain of salt. With that said, ChatGPT is 100% right about all the nice things it said about me. I’m amazing — zero notes from my end.

Seriously though, all of the aspects of the story it called out were the ones I felt most confident about — to put it a different way, I might have been offended if ChatGPT said, “I don't like your concept, tone, world-building, and imagery in this.”

With this section, ChatGPT affirmed that the things I thought I got right were actually solid. If I pick this story back up, I'll know what to keep consistent — all thanks to my new, reliable, potentially dystopian-future-responsible peer editor.

Areas for Improvement:

  1. Plaque Description: The critique of the plaque's wording is funny, but it might be tightened to enhance its impact. Perhaps focus more on the irony and less on the semantics.
  2. Information Density: While the exposition is mostly well-handled, there are moments where it feels a bit heavy. For example, the detailed explanation of werewolf characteristics and the societal debate could be spaced out or integrated more naturally into the narrative.
  3. Character Introduction: Ronnie's physical description and background as an ex-marine are important, but consider weaving these details into the story more subtly rather than presenting them in a block of text.
  4. Flow and Pacing: The transition between the office description and Ronnie’s personal details feels slightly abrupt. A smoother transition would help maintain the flow.

Do I agree with ChatGPT?

I mostly agree with ChatGPT here, but I don‘t know that I really have a choice. It’s always tough to look at thoughtful, measured, constructive criticism and say, “Nuh-uh — I totally did everything right.”

I don't all-the-way agree with the “plaque description” critique. It might just be me, but I thought it was fun — I liked the whole, “The office was like a family…” part. Still, I could see where ChatGPT was coming from. I think I have a tendency to beat readers over the head with my writing sometimes. That might have been the case there.

I also get what it meant when it called out the “information density” aspect. That's one of the big issues I feel like I always run into with my creative writing (see my beating readers over the head comment from like three sentences ago).

I can get nervous and wind up writing with “Did you get the joke?” energy. It was helpful for ChatGPT to mention that and provide solid direction about it.

As far as the “character introduction” element goes, I think ChatGPT has a point. The way I introduce Ronnie is pretty jarring — and that trend carries over to the following point ChatGPT made.

All told, ChatGPT makes some solid arguments here, and I agree with most of them.

Specific Revisions:

  • Plaque Commentary “A Force. A Friendship. A Family.” The gold plaque above the Silver Bullet Enterprises entryway always irked Ronnie. For a company named Silver Bullet, why gold? And “friendship” in an office? It seemed like they were just trying to fit in another “F.”
  • Exposition Breakup: It was a tough gig — werewolf extermination. Dangerous, thankless, and controversial. But you got to kill something big without legal recourse, a perk for many in the industry. You might expect mystique or magic at a werewolf extermination office — cauldrons, spell books, witches. Silver Bullet's central Massachusetts office, however, could be any generic midsize firm: slate gray cubicles, stale coffee scent, creaky chairs.
  • Integrate Background: Ronnie, an ex-marine of 5‘7 with a jawline he’d been complimented on exactly twice, pushed through the double doors. He'd grown used to the protesters’ taunts — today’s insults were mild. His slender frame and “kind eyes” (as the bartender at CB Hannigans once called them) belied his tough exterior.

Do I agree with ChatGPT?

I agree with ChatGPT to a certain extent here. I think the sections it called out for revision do need work, but I wasn‘t in love with ChatGPT’s suggestions — especially with the “plaque commentary.” It omitted what I thought was the most interesting part of the description (the “family” bit), and I felt like ChatGPT's suggestion was a bit too hokey.

Still, until this article is published, no one but ChatGPT and I (and Next in AI's fabulous editor, Martina Bretous) will have seen this story. Maybe, I can get a second, non-robot opinion on the plaque once this piece goes live.

I did think the next two suggestions were pretty solid. They condensed some over-the-top, excessively exposition-y parts without sacrificing too much. I don‘t know that I would use ChatGPT’s exact language here, but it definitely set me in the right direction.

ChatGPT also offered me a concise overall impression.

ChatGPT's Overall Impression:

Your story has a compelling premise and an engaging voice. With a bit of tightening and smoother integration of background details, it will read even more fluidly. Keep up the good work!

Do I agree with ChatGPT?

As I said, I‘ll take any compliment on my writing I can get, so ChatGPT gets some immediate points in my book — but beyond that, it was surprisingly thoughtful and empathetic with its critiques. That’s pretty good for something that's going to destroy humanity as soon as it becomes self-aware.

Is ChatGPT a valuable creative feedback resource?

The short answer? Honestly, yeah.

I think the insight it produced provided a pretty valuable jumping-off point for the rest of the story. It confirmed some issues I thought I was having (particularly, too much exposition) — and the advice it offered was specific and actionable. I'm also sure I could get more detailed takes on its opinions if I prompted it to elaborate point-by-point.

I didn‘t love all of ChatGPT’s revisions, but they shed light on some valid issues — getting a little lost describing the plaque, overdoing Ronnie‘s physical description, and letting excessive world-building get in the way of the story’s flow. Also, all of the feedback was conveyed firmly but compassionately.

Ultimately, this experience was simultaneously helpful and terrifying. There's a scene in the movie I, Robot where Will Smith asks one of the now-sentient robots if a machine can write a symphony — essentially whether AI has the capacity to replicate human creativity.

If this is any indication, AI is pretty much there.

ChatGPT's feedback here might have seemed a little flawed and hastily strung together, but again, it produced all of that in less than 30 seconds. A robot immediately turned around a series of fairly detailed, constructive notes that I plan on incorporating as I work on this story more.

That's equal parts amazing and unsettling.

Click Here to Subscribe to HubSpot's AI Newsletter

Related Articles

Pop up for SUBSCRIBE TO HUBSPOT'S NEXT IN AI NEWSLETTER AI NEWS IN YOUR INBOX

A weekly newsletter covering AI and business.

SUBSCRIBE
Pop up for SUBSCRIBE TO HUBSPOT'S NEXT IN AI NEWSLETTER AI NEWS IN YOUR INBOX

Marketing software that helps you drive revenue, save time and resources, and measure and optimize your investments — all on one easy-to-use platform

START FREE OR GET A DEMO