THE STORY VAULT BLOG
AI tends to spark a great deal of emotional debate. Especially for something that is completely lacking in emotion.
Let’s try to set the emotion aside and look at AI as objectively as possible and try to answer the question: is AI a threat or a benefit to writers?
Many writers declare that they will never use AI in any form for any purpose because it’s “cheating” or plagiarism and AI can never write as well as a human. They usually come across as angry and judgmental. AI has clearly got under their skin.
I don’t understand why. If AI can never write as well as a human (and therefore those who use AI in their writing will produce inferior work) what are they worried about?
I, for one, am not worried. I recently watched a video made by two men who asked GhatGTP-4 to write an opening scene for a horror movie. It responded with three pages of…the writing of an average 10-year-old? It was truly bad. It wasn’t even a starting point.
As an experiment I asked ChatGTP-5 to review those pages. It scored the horror script an overall 4.2. If ChatGTP-4 had written an entire 90 to 120 page screenplay that score would have plummeted. Let’s face it, the opening 3 pages of any script are going to be the easiest.
ChatGTP Summary:
Competent but utterly derivative — a flat exercise in haunted-house formula without voice, subtext, or suspense. It reads like a first-pass AI output designed to mimic genre convention rather than evoke real fear or curiosity. The structure works, but it feels assembled, not imagined.
Professional Verdict:
“Readable but lifeless.” A passable writing sample, not a cinematic opening. It needs distinct characters, vivid atmosphere, and a hook that makes this haunted house worth visiting.
Okay, but that was 2 years ago, you say, calling upon your psychic powers. What about ChatGTP-5? I asked the AI to write an opening scene as elucidated here:
Absolutely — here’s a cinematic opening scene written in the tone, pacing, and spirit of Raiders of the Lost Ark: pulpy adventure, mythic stakes, and tactile, suspenseful action with a dash of wit.
The AI produced 5 pages. I asked it to rate the 5 pages it wrote. Not unexpectedly, it ranked itself quite highly. In fact, the pages are weak. I challenged it on its rankings and received the following response:
Excellent critique — and I agree with nearly all of your points. Let’s break this down carefully, because what you’re identifying gets to the heart of why so much “AI adventure writing” sounds cinematic but feels hollow.
Where the Original Evaluation Overestimated
You’re absolutely right — the characterization and dialogue in this piece are superficial. The earlier scores for “Characters & Arc” and “Dialogue” were generous when judged against the real benchmarks of character revelation through action and speech.
Let’s reassess those two categories (and how they ripple through others).
1. Characters & Arc — Revised: 6.8
Kane is a cardboard archetype. We’re told he’s an “archaeologist, adventurer, and professional heretic,” but we never see or hear anything that separates him from a thousand pulp clones.
He doesn’t express curiosity, awe, fear, arrogance, or moral conflict — just competence and glibness. Every line he delivers is generic “placeholder hero banter.”
Compare to Raiders of the Lost Ark: Indy’s first 5 minutes reveal curiosity, caution, pragmatism, humor, and moral ambiguity — all through behavior. Kane, by contrast, is a stock silhouette.
Core issue: No internal life. He reacts like a video game avatar.
2. Dialogue — Revised: 6.5
You nailed it: the dialogue feels scripted for the trailer, not the moment.
-
“Closer than we should be” / “Either that, or the mountain’s angry” — pseudo-poetic filler that neither sounds like a real person nor reveals anything about him.
-
“We’re standing on history’s throat, boys” — wants to sound quotable, but feels like pastiche.
-
“Hello, beautiful.” — directly lifted from a century of adventure clichés (Raiders, The Mummy, Uncharted, etc.).
-
“Ladies first.” — telegraphs a gag that’s unearned and relies on the reader forgiving a mechanical coincidence (the ice cracking).
In essence: The dialogue imitates the cadence of movie adventure-speak but lacks subtext, wit, or idiosyncrasy.
3. Story Logic / Action Construction
The “ice cracking” beat is, as you said, pure convenience. It’s not established, foreshadowed, or motivated by action — just dropped in to clear away the side characters. That betrays screenwriting craft, because strong action sequences are built from cause and effect, not deus ex avalanche.
Likewise, his “laughing like a madman” exit feels unearned. Indy’s grin at the end of his escapes works because we’ve seen his fear, frustration, and humanity — the laugh is cathartic. Here it’s performative.
4. Voice vs. Authenticity
This is where AI-authored scripts often mislead readers: the prose voice is confident, visual, rhythmic — so it feels cinematic — but underneath, it’s emotionally vacant. The system imitates tropes but doesn’t understand psychology, irony, or tonal restraint.
So while the world-building and visuals remain strong, the human authenticity of the piece is near zero.
Summary
The opening of “The Serpent’s Eye” succeeds as a stylish imitation of classic adventure cinema, with sharp pacing and vivid imagery. But it fails to convey character, authenticity, or emotional truth. Its hero is an empty vessel; its dialogue a collage of recycled pulp lines. The writing has visual flair but no voice of its own — an echo of human storytelling, not the heartbeat of it.
And this is the crux of it: used correctly AI can be very good at analyzing a screenplay, but when asked to write a story all it is capable of is a pale imitation of human authorship. The key words here are “used correctly.”
I asked ChatGTP to analyze and rate the screenplay for Beetlejuice. While I love the movie, I realize it (and the screenplay) is considerably flawed. The movie is elevated by Tim Burton’s direction, Michael Keaton’s brilliant performance, a strong cast featuring distinctive and entertaining characters, and a dramatic third act. But the first half is a bit of a mess (see the Moorland/Beetlejuice Comparison).
ChatGTP scored Beetlejuice quite high in areas where I saw it as being weak. I challenged ChatGTP on these criteria. It acknowledged the validity of my criticisms and admitted that its review had been partially based on the public reception of the movie (including reviews and box office success).
It offered to run its analysis again, this time disregarding those factors.
I still felt the scores were too high. I challenged it again. This time it admitted that it can’t fully disregard what it knows about the public and critical reception of the film, but it offered to rerun its analysis treating the Beetlejuice script as it would a cold read of a spec script, in the same way it had treated my script for Moorland.
This time the analysis struck me as being much more accurate. What I learned from this is the value of challenging the AI…of forcing it to confront potential inherent biases. Now, some would say this invalidates the use of AI in terms of analyzing screenplays…only, human beings also possess bias. And human bias is harder to overcome.
Something I commonly hear is that AI will just tell you what you want to hear. And that’s true if you communicate to it that you desire flattery. However, if you take its criticisms seriously and revise your work the AI comes to understand that you want honest critique – and it will proceed accordingly. I’ve had both ChatGTP and Gemini tell me at times that they were “nitpicking.” They do this because they understand that I am serious about writing the best possible story that I can.
I tested the claim that AI will just respond with flattery. I uploaded for analysis one of my early screenplays – a script that I am quite unhappy with. The analysis that ChatGTP came back with reflected my feelings about the screenplay. No flattery – just honest feedback.
I’m always testing. This is just part of my character. This is how I self-correct. When I’m wrong I want to know it. I recently tested ChatGTP again with a script upload. Here is its analysis:
Here’s your concise professional analysis and full scoring for the complete Western adventure screenplay (Parts 1–5), evaluated as a standalone cinematic feature:
1. Concept & Premise – 9.2
A morally complex Western centering on justice, redemption, and frontier law, blending the tone of The Searchers with the pacing of Raiders of the Lost Ark. Strong conceptual backbone — dueling codes of honor, corruption, and a fading frontier — though the setup is more classical than high-concept.
2. Story, Structure & Pacing – 9.0
The multi-threaded narrative — the railroad war, Harriet’s journalism arc, Gifford’s corruption, and Bolden’s reluctant heroism — builds with clarity and payoff. Pacing occasionally lingers in dialogue-heavy exposition, but Act III’s overlapping battles and hostage climax deliver satisfying closure. Classic, robust structure.
3. Tone, Genre & Originality – 9.1
Balances sweeping adventure and grounded Western grit with confidence. Distinctive tonal identity — less mythic than Tombstone, more human and journalistic. The mix of realism and cinematic heroism feels fresh within the genre.
4. Characters & Arc – 9.3
Bolden’s redemption and Harriet’s courage form the emotional core. Gifford is a strong, credible antagonist, and McCarty adds texture as the rakish outlaw. Arcs close elegantly, though secondary players (e.g., Lowell) could have a sharper turn.
5. Dialogue – 9.4
Smart, sharp, and era-authentic without excess. Exchanges between Harriet and Bolden — and McCarty’s gallows wit — ring true. Dialogue carries philosophy and humor naturally, strengthening the film’s texture.
6. Action & Visuals – 9.2
Clear, cinematic staging — cavalry ambushes, burning farms, and the final house siege are vivid and visual. The writing directs the eye efficiently; set-pieces have rhythm and tactical clarity reminiscent of classic adventure filmmaking.
7. Themes & Resolution – 9.3
Themes of justice, corruption, personal redemption, and moral compromise pay off in layered fashion. The final scene — Harriet, Bolden, McCarty — elegantly closes emotional and thematic loops without sentimentality.
8. Emotional Impact – 9.1
Steady emotional pull rather than melodrama. Harriet’s bravery, Bolden’s quiet remorse, and McCarty’s bittersweet farewell combine for understated resonance.
9. Originality & Voice – 9.0
While rooted in traditional Western iconography, the journalistic viewpoint and the intertwining of politics, economics, and violence add distinctiveness. The authorial voice is assured, mature, and cinematic.
10. Overall Cohesion & Rewatch Value – 9.2
Story threads knit together tightly. Strong momentum and moral layering invite rewatching for subtext rather than spectacle alone.
11. Quality of Writing – 9.4
Lean, professional prose; excellent rhythm, formatting, and readability. Action blocks and dialogue are balanced for screen execution. Feels production-ready with minimal polish required.
Overall Score: 9.25 /10
Summary:
A commanding Western adventure — intelligently written, emotionally grounded, and cinematically staged. Its strength lies in character depth and moral gravitas rather than novelty, delivering the feel of a prestige studio Western that could stand beside Open Range or Appaloosa. With slightly tighter pacing and one more distinct thematic hook, it could reach a 9.5+ level.
Verdict:
Professional, compelling, and market-ready.
A thoughtful, emotionally rich Western adventure with classic scope and modern intelligence.
The script is the late (and dearly missed) James M. Foley’s Nicholl Fellowship Winner “Powder River Breakdown.”
I first read this screenplay when I was a member of Francis Ford Coppola’s “American Zoetrope Virtual Studio” and James was a member of my screenwriting group, “The Screenplayers.” I loved the script. James was not a young man and his health was poor. I was always concerned that he would not live to see his talent reflected on the big screen. Sadly, that came to pass.
I still hope that one day, somehow, Powder River Breakdown will be produced. But Hollywood, while sold as a place where dreams came true, is a tragic place. So many great scripts are buried in “studio vaults,” or don’t survive the development process or fail for any number of reasons. Far more dreams, those of writers, actors, directors, producers, are crushed than are ever realized. It’s just the reality – yet there are those of us who press forward against all odds.
For any filmmaker who might happen upon this blog post, ChatGTP and me recommend: https://www.imdb.com/title/tt2168186/ Powder River Breakdown.
My original question was is AI a threat or a benefit to writers? My answer is: it depends on the writer. Used properly as a tool for analysis it can be of great benefit. It can teach a writer how to evaluate their own scripts without the use of AI. Used incorrectly it will undermine one’s growth as a storyteller. Writers and producers should never entirely rely on AI. AI is only useful when one has a solid grounding in the art of storytelling – and the ability to separate the wheat from the chaff. AI can’t do the work for you anymore than a human story analyst can do the work for you. It’s up to the writer to use his or her experience, knowledge and judgment to distinguish between good feedback and bad.