On Writing : Using ChatGPT for screenplay feedback?? by Lynelle Paulick

I wouldn't trust AI to give me feedback,Lynelle Paulick. I've heard AI gives weird feedback.

I used ChatGPT 4 once for coverage. It told me my protagonist didn't arc. My protagonist literally starts as a diagnosed agoraphobe and ends by rescuing everyone in New York City,

Ohhhhh, I LOVE both of these!! hahahaha. Alrightee right, my decision may be made! :) Thanks, Maurice and Michael.

You're welcome, Lynelle Paulick.

I think that using AI as a tool for research, organizing thoughts, etc. is one thing. But when it comes to something as important as feedback or notes, you want to make sure it's coming from a real person who can connect with your story and characters on an emotional level or pick up on nuance.

Beautiful. Thank you, Ashley. This and the other responses are making it pretty clear to me where not to go at this stage! ;):)

Hello CHATGPT is great software, a lot of creatives are using it, I am currently writing a historical thriller, and using it with certain scenes has helped, think of the film industry, they use AI to help them. I would never see it to write the whole book, so if it offers feedback go for it

I am not certain how much of it is feedback and how much of it is just a computer spitting back at you the same information that you gave it.

I've actually gotten some VERY insightful notes from Chat GPT....

Chatgpt is capable of handling basic information and offering feedback, albeit limited. It ties up limited summaries in a pretty bow and skin-deep recommendations. Good start, but if you want something more robust, your best bet is to talk to a human.

Some great notes here already, so I'll echo - for brainstorming and researching, but certainly not for analysis of one's work.

you can ask PI, GPT, and Gem to act or roll play as a screen writer and it will do it just like one. I have more of these.

What incredible responses [jawdrop emoji]. Thank you, all.

It's a great tool to help with loglines, especially when you struggle with sharing a story in a single sentence.

I'll also echo the same thought as many: while I'm generally against AI in this field of endeavour, if you're doing it for some basic research or to inspire you (maybe creating some visuals to help put a character firmly in your mind's eye when you write), then that's one thing, but for analysis, it's a big fat no. You might think what you get from AI notes is insightful, but I can assure you any script consultant or story analyst worth their salt will find holes. And at the end of the day, who do you make your stories for? Humans or AI? The answer to that question will inform you who you'll really need to check and provide feedback on your story.

I have read some AI generated novels and they are awful. As an English professor, it concerns me because some people are actually reading them. I'm finding that AI is great for generating scenarios. As I move through the first draft of my fifth novel, I'll reach a "what-to-do-next" point. I feed in some ideas on AI and see what it comes up with. I do that repeatedly. It's like adding fertilizer to a garden. The garden in my creative mind.

John! That is So interesting. Certainly a reasonable and controlled experiment you are doing -- and: with the awareness you have, you're not ever going to fall "victim" to it. I really appreciate this comment, thank you.

Lynelle: Thanks for your comment. I got into the writing game late so I am not disillusioned about a "career." I do have four novels on Amazon, traditionally published, but I'm not getting rich. Unfortunately today, the market is dead for you if you are unknown. I've been told I need to get involved with internet media, but that is a swamp that can consume your time. I prefer to hone my craft. I have several friends that are deeply involved in the internet and they spend many hours working at it. They cannot tell me the percentage of their success. If you are on the internet, you're not writing. My motto is BICHOK: butt in chair, hands on keyboard. If you don't write it down, it doesn't exist, and great novels are not produced in our mind alone. As I tell my English students, if you want to learn it, write it down.

I used AI to give me critiques of loglines, then asked which of my screenplay ideas were the "best." Amazingly, it gave me an order that I agreed with and good reasons for the choices.

I Worte a story using a one dice RPG system and it came out ok.

John, this is fabulous. All I can possibly say is YES.

Thanks, Stan, that is good to know!

Preston, I may well take advantage -- MUCH appreciated.

I edit books on AI and this question come up a lot. ChatGPT is a large language model, not really intelligence. It can mimic language, but its intelligence, accuracy, and ability to analyze is limited.... for now. Great strides are being made in the large language model area and I don't doubt we will see systems that are able to accurately analyze screenplays in the next few years.

For now though, I wouldn't trust its accuracy "out of the box" at evaluating any creative work.

Chiming in to say -- Leaning HEAVILY towards a realllllllly bad idea. Your audience isn't an arbitrarily programmed computer model. Your audience consists of people (who are the ones at the end of the day will or will not pay money to watch whatever your create). Get people to read your projects. Stage 32, Reddit, Discord, Twitter, maybe even Instagram all have communities you can easily integrate into to meet other writers and exchange ideas. Give it a go, put yourself out there and do a manuscript swap for feedback. Learn to give notes and receive them. It's a super rewarding process. I would just never trust AI to reasonably evaluate any creative material. I personally write as a form of artistic expression and I can't imagine a world where the emotion/feeling part of any story isn't entirely lost on AI

Dustin and Pat -- crissake, I'm also quite naive about how to actually Answer a specific person on Stage 32. I'll get there. But at this moment, I want to thank BOTH of you for the helpful and definitely insightful input. I love all you people, it's amazing!!!

Although I may be biased-- since I'm a longtime freelance screenplay analyst in addition to being a working screenwriter-- and have no personal experience with seeing a script evaluation written by AI, I don't think that AI can evaluate a screenplay effectively from a creative standpoint-- and provide actionable, specific advice on how to revise, which for screenwriters is (or should be) the purpose of getting a script evaluated in the first place. In evaluating a script, I not only have in my memory banks the history of all the films I've seen (from silent movies, to the present) and the thousands of screenplays I've evaluated for screenwriters, movie studios, film agencies, etc., but also my experiences as a screenwriter, a knowledge of history, current events and cultural context. And also I take into consideration the writer him or herself, their background, and difficulty factors in any changes I'm recommending. As any screenplay analyst will tell you, a huge part of the job is psychology-- understanding the emotions and behavior of human beings: the characters in the script itself, and the writer of the script. Writers can often benefit as well from hearing from more experienced screenwriters the fact that writing is difficult and that there's nothing wrong with them if they are struggling. So, knowing when and how to empathize, and when the writer needs a gentle kick in the pants, is also helpful. AI may be a useful tool for writers at the "idea" stage, especially if they are aiming to write a mainstream Hollywood studio film. It may be helpful (and probably is) for studios to use in trying to predict box office numbers, etc. But I can't imagine how AI would be useful to writers in getting their script evaluated if they hope to improve it, and before giving my script to any person, contest, or company for evaluation, I would ask if they use AI at any stage to evaluate scripts.

Or...to make a long story short, I agree with Pat and Matthew:-)

AI feedback is not real feedback. It is simply reflecting information it already has based on learned information. Do you realise whatever information you enter into an AI platform that information is absorbed into its system to learn from and it will be used to answer future questions? If you feed an original idea into an AI platform it is no longer an original idea, it is integrated into their knowledge bank and could be spat out to answer another question at any point in the future. I would be careful IMHO.

Thanks, Matthew, Staton, and Matt. Huge knowledge and input here! I do wonder what the answer to this general question will sound like as AI technology continues to evolve (I hope it doesn't, but assuming it does...).

I think the best feedback you'll get is a brain with blood running through it. AI is a machine that can only think on a machine level what was put in it. It can't think like a human being. I'd read what you have for you but I need to write a bible. Good luck with your query.

Very, very bad idea.

I can't speak to ChatGPT 4 - but the free version would be useless as you basically have to write out the whole plot in the text prompt - it's incapable of reading your script. The main problem with chatbots is since Large Language Models are predictive based on math, it doesn't really help with artistic creations.

William Joseph Hill That's not true. I do not have 4. I only use the free version. And it reads and gives feedback to my scripts just fine. I copy/paste the material 10 pages at a time, and it's able to "remember" all of it to give overall notes in addition to it already giving notes on each 10 pages.

AI, at the moment, is the most basic version it will ever be because we are at the beginning, it will only get better in terms of output. Whatever you add as a user today will be used by the AI platform to learn and at some point it will regurgitate your information for someone else. That's how LLM's work. I would not be putting any of my original work into an AI platform to gain feedback. IMHO.

Yes, I was just thinking about this. ChatGPT collects people's data for the AI models to learn and improve. I would be careful sharing my work with AI.

I took an AI certification course with Ragan Communications and saw that it’s all smoke and mirrors. For example, data shows LLMs are already starting to “plateau” and generate output based on themselves and other LLMs, leading to less quality results. https://m.youtube.com/watch?v=dDUC-LqVrPU

In the class, even when you buy an enterprise license, they said you cannot guarantee your proprietary data would not feed LLMs given to other people.

Moreover, if it’s collating “ideas” for you from a million seasons of Law & Order or every frame of Star Wars, do you want to open yourself up to the legal liability or notations of derivative nature?

Generative AI is a bad thing before you get to the environmental costs, before you get to the employee treatment, before you get to the outright theft of IP and likenesses.

Travis Seppala - Oh yes, that is true, if you break it up you can get feedback from Chat GPT. What I meant to say was that it can't read a PDF of your script - but yes, copying and pasting will work. Do you find the Ai feedback to be useful?

William Joseph Hill Much like human feedback, I find some of it helpful (stuff that's insightful, pointing out issues I never would have considered along with reasons) and some that is not (vague generalities, parroting obvious story beats back to me, making suggestions to "fix" things that would wildly alter the story I'm trying to tell).

My general question to anyone considering using any form of AI to get screenplay analysis-- whether through Chat GPT, or a screenwriting service that (God forbid) uses AI to evaluate scripts instead of human script readers-- is, why? Because it's often free or costs less than hiring a qualified human being to do this? That would be about as useful (and safe) as entering one's symptoms into Chat GPT for diagnosis and treatment of a medical condition, instead of seeing an MD.

Suuuch a great question, Staton. And yeah, it points right back at me, the originator of this thread. This incredible thread, I should add. It was the idea of a well-intentioned and completely naive third party who was just getting into a flirting match with this new and "fascinating" ChatGPT app, an innocent-enough bit of advice simply intended to direct a couple of clueless first-time screenwriters in their struggle to find humans who would 1) be willing to read the draft for free; 2) be someone we could trust not to go share it on the street inappropriately; and 3) be someone we actually know personally who would be qualified to give advice and direction that we might be able to use. My writing partner was 100% opposed right from the start. I -- well, I just got onto Stage 32 -- and here we are. :) I've learned a massive amount in this discussion.

Thanks for your kind and receptive response, Lynelle, much appreciated. Script notes one gets free or for low pay from civilians (human or otherwise) can be very hit-or-miss, and if you don't know how to filter conflicting or bad advice, you may waste a lot of time revising your script in the wrong direction. Just to mention that If you register your script for copyright with the Library of Congress you are well protected, except for your story concept/idea, which is not covered by copyright. Scripts and ideas rarely get "stolen", though-- it's really the execution of the idea that counts. And if you work with pro script readers, it's part of their job description that all scripts they evaluate are kept confidential. If your script is so great that people might want to steal it, I'd take that as an enormous sign of encouragement:-)

It all comes down to money and time for me. The prices for consultations are not cheap. I don't know the reviewer from Adam. I don't trust the reader is giving my script the attention it needs as they have a slush pile of scripts to get through and by the time they get back to me, I have already changed the script, so whatever critiques they have more than likely don't matter because I have changed the script. So, it's a moot point and I have wasted not only my own time and money but someone else's time as well. I am sure the reviewer is thorough and provides better feedback, but it comes with a cost.

I have to concur with the main reason being cost -- many writers simply can't afford paid coverage, and the free feedback can be useful in the early stages of script development. There are also writing groups that you can get involved with to get feedback from other writers, which will be more qualified feedback, and can probably point directly to what your script needs, instead of generalities. At some point, when it's ready to be pitched, paid coverage can be a good investment. But I think it's being resourceful to find low cost or free tools that can help. I've found that there's no shortcut, so the secret is being able to cultivate a sustainable process so you can keep plugging away.

I just want to jump back in and thank you all for having such a respectful conversation about a difficult subject matter. Conversations like this one showcase what's so special about this community. I appreciate all of you for managing to share your opinions with respect to each other's points of view.

Meta will be using your FB and Instra posts for AI. https://www.abc.net.au/news/2024-06-10/instagram-facebook-train-meta-ai-...

I use it to enhance my loglines too Lynelle Paulick and never used it for covereage.

There are many very good comments on this post. But, whether we like it or not, AI is here to stay.

I think AI is a juggernaut that will roll over everything in its path. I read my past comments and I don't want anyone to think I in anyway understand how to use AI. As far as writing goes, I am a dinosaur who starts fire with sticks. I am floundering with all this knew stuff. My youngest is a tech support person and I am always screaming "help." Most of the time I just sit down and write. I enjoy the process of writing and editing. Everything else is just sauce, and in some cases, it ruins the entre.

If any of you are looking to ask more questions about AI directly to a professional best equipped to answer them, you should jump into TODAY'S Ask Me Anything in the Writer's Room Lounge. This is an opportunity to be better informed about how it's affecting the industry, how it can be used responsibly, what the common misconceptions are around it, what the practical changes and concerns are, etc.

Joshua Young, Screenwriter, Filmmaker, Video & Animatic Editor, and AI Consultant (Netflix, PBS, OWN, TLC, etc.) is sharing his time with us to answer YOUR questions.

The post is live for you to comment with your questions NOW!

https://www.stage32.com/lounge/vip/Ask-Me-Anything-AMA-Friday-6-21-to-Sa...

If you're not already a Stage 32 Writer's Room member, you can sign up for your first month entirely free so that you can join in on these opportunities.

https://www.stage32.com/writers-room/plans-vip

Well Said John Roane

I asked my English class last semester if anyone had read a book recently. No hands went up. Then I restated my question. Did anyone like to read a book, just read for fun, specifically a novel? Still no hands came up. My students were twenty five recent high school graduates. I talked about journaling and the joy of sitting down at the keyboard or with a pen and paper and enjoying the process of creating a narrative fresh from their own brain. Editing of work completed was an alien concept. I hate to use the phrase "dumbing it down," but as the semester progressed, I gradually simplified my lecture material to the point that it was, for me, at the level of a middle school student. I was trying to teach a card trick to a bunch of chickens. I am not a teacher. I am an entertainer battling to hold the attention of an uncommitted and uninvolved audience. It is all very depressing and forces me to reconsider the goal of my writing efforts. What is the future of the novel and why am I creating them? Am I a purist, producing narratives simply for the joy and satisfaction of laboring in the vineyard of solitude? Following on from that, what has this to do with AI? If my students can't read and cannot put together a coherent and grammatically correct sentence, how can they use AI, this complex new tool I do not understand? The college is concerned they will write all their papers using AI. Am I the dumb one here?

John Roane Sorry to hear that, I do try to get my writing up to adult reader level being that I never really had a great education having to go to special ED classes for most of my life, I grew up with a speech impairment more do to my tongue was too big for my mouth. But after trying to work with writers on comics who turned out to be real jerks I looked into writing my own or rewriting public domain stories and at this time I fear that my writing will never go above YA in writing style. I have some of my writing up here for people to read, I use AI for parts that I have issues or troubles with, but I try to write my stories my own way in my own voice. I do and would like for you and others to read and comment on my stories. I want to get better but getting help from someone who charges a lot for that type of service even here makes me more or less go with AI, I wouldn't mind trading my time to read for someone here if they would read for me and tell me what they think and what I think. So in the end just like with any other word processor program, AI is more a tool than a cure-all item.

There is a line when use becomes abuse. The calculator was a cutting-edge technology that could automate long tedious tasks. In the 1970s, people started getting stupid, getting so good with the calculator that some math nerd got a chuckle when he learned that 0.7734 when inverted was "hello." 5318008, which when turned over spells "BOOBIES". Point I, the tool is only as smart as the user. But, it can be dumber, for example, if one were to ask ChatGPT to write a sitcom script, it wouldn't be able to make the script funny because comedy is subjective. If, however, one were to request feedback on a script, it could compare it to scripts that are similar and provide feedback based on this information. The creative liberty and onus still falls back on the creator to create interesting characters with sharp dialogue, but instead of hitting walls and procrastinating, the creator could chart his way through the script, potentially making it better, having critiques from ChatGPT. But, like with everything else, user beware. AI is definitely not a cure-all. It won't make you magically become a writer or a mathematical genius or an expert on anything. And, it will never replace the human mind because the human mind is constantly expanding faster than any AI tech can. Because we are working towards something. We are always evolving. Tech can only mimic life - not experience it. The tools a craftsman uses are only as good as the craftsman himself.

John, that is STUNNING. I had my hand over my mouth as I read it. Wow, nothing to say...I'm afraid what you're describing is Not the exception...

So Kevin! Your comments here are So entertaining :) ! But this: All that keeps coming up for me when hearing arguments about the inability of AI to replace the human mind is that over and over, more and more, I am reading these narratives (for instance, a recent interview I saw on X with a fellow named Geoffrey Hinton [https://t.co/11A7fJpLcs]) that postulate -- or even literally promise -- that the long-term potential of AI is to rendor absolutely unnecessary the human mind and its input. Which means that every argument we have here "could" someday be literally answered by some AI entity itself. You know...the Hal story come to life. Does any of this bother Anyone? Is it BS to even concern oneself with this fantasy-like "impossibility"?

The evolution of AI is going at a snail's pace. Five years ago, no one ever heard of ChatGPT. Two years ago, someone somewhere became fearful of ChatGPT. Thought it was magic. Whatever. We fear what we don't understand. The assertion that AI is going to rule the world is something out of a cheesy spy thriller. Nothing more.. but, because of this scare tactic by someone behind the screen, media, YouTubers, and others are throwing grenades, inventing nothing but speculative drivel about some doomsday scenario where humans become submissive to artificial intelligence. I wouldn't call AI inferior, but it's really no match for anyone with a college degree. And, I believe it will remain that way. There is no evidence to suggest otherwise. Even if there comes a day when AI surpasses expectations, we will probably be long gone, so isn't it kinda a waste of energy to fret about this scenario? Anyway, I'm going back to writing.

That is why Disney is spending millions on Ai companies & development right now as we speak, because they do not want to get left behind in this new revolution taking the world by storm:

Disney announced on Wednesday that it’s making a strategic investment in five AI and VR startups as part of its 10th Accelerator Program. Notably, three of these startups – ElevenLabs, PrometheanAI, and AudioShake – are focused on leveraging artificial intelligence to drive innovation in their respective fields.

And is this why Disney is laying off thousands of jobs, especially Pixar?

https://techstartups.com/2024/02/23/disney-makes-major-investments-in-ai...

Gregory: I detect an apologetic tone in your narrative. Don't ever apologize. I think I have written a few good pages, but if I had a nickel for every page of garbage I wrote, I would be a wealthy man. The fact you are writing is the mark of success. The mark of a good writer is never being satisfied with our work. That is the source of irritation with my students. They produce their initial draft and are satisfied with it. Drives me crazy. They turn in papers full of minor and major grammatical errors that are glaringly obvious. The secret to good work is the struggle of editing.

John Roane Thanks for the kind words, I can be a bit self-conscious and hard on myself. I sometimes feel like my writing will never be good enough and I second-guess myself too. I use spell check on the word processor programs I use for my writing. I find it odd that your students don't use that. I have a few that I use and know of a few. Sometimes I think I'm doing something wrong and find I was doing it right! I still have things I need to work on with my writing. being self-taught is not always easy.

Congrats on your "small" victory Mariia Rudenko. It's a portrait of an extraordinary artist

Actually, Lynelle, I've decided to give AI my logline and 3-paragraph synopsis and let her write my screenplay. Meanwhile I'll grab a coffee and watch TV. Now that's a career I can aspire to! :)

John Connell , already been done, but failed to grab an audience. The first feature length film in a cinema, screenplay written by AI, didn't happen in London last week - read why - https://www.bbc.com/news/articles/cjll3w15j0yo

John: I can identify 100%. I have several novels, short stories on Amazon(2 different publishers) - I go out and do readings and signings. But here's the trick: if you're not on TV that morning, no one will buy your book. There's a local former TV anchor who did a book signing. He sold 88 books in an hour. In 2021 I started a Memoir Writing Workshop on zoom. It still is going. I got them published. In this business, it's who you know.....

This post really stirs the pot - doesn't it? Most of us are on one side of the fence or the other regarding AI. Even though I am also an engineer - I am firmly against AI for any creative endeavors. So much so, when I found out the graphic artist I hired used AI for the first book in my fantasy trilogy, I hired a different graphic artist to redo the cover with no AI. My graphic design costs doubled, but I felt it important to stand for my beliefs. None of my work (book covers for my books, my manuscripts, and audio books) will utilize AI in its creation. Amazon is offering to select authors an AI-driven opportunity for free Audible creation. I spent $1500 on the Audible for my science fiction novel so I can't tell you how tempting that was - but I said no thanks. I get it - AI allows us to do some things we may not be able to afford otherwise - but for me (and I was recently laid off, so my budget is always a priority) I have to say no to AI. I will wait for an audio book for my fantasy book for if/when I can afford it. As a creative - I challenge all of us to reject anything AI driven. We need to do the hard work and collaborate with one another - not allow a computer algorithm to take the place of our humanity and our creative souls. I have worked in IT, and I can tell you with certainty that they are proceeding ahead with AI solutions because it is trending and bringing in the money - but no one understands the ethical consequences that will result from this irresponsible behavior. The CEO of the Authors Guild, Mary Rasenberger, was recently published in the LA Times and writes, “We cannot trust tech companies that swear their innovations are so important that they do not need to pay for one of the main ingredients — other people’s creative works. The “better future” we are being sold by OpenAI and others is, in fact, a dystopia. It’s time for creative professionals to stand together, demand what we are owed, and determine our own futures.” If you haven't thought about the ramifications of AI on the creativity community, I hope you will do so. All of our futures depend on being in solidarity with one another as we strive to provide what only we can uniquely produce as artists/creatives - to a world in desperate need of entertainment (film, books, art, music) that can provide a welcome distraction from these troubling times.

Copyright applications must disclose whether the application includes AI-generated material and the extent of its use. However, a writer could modify material originally generated by AI to such a degree that it meets the standard for protection. It will be difficult to distinguish if a writer has used Plotdot and Nolan AI in their screenwriting process. This would be applicable for entering screenwriting contests as well.

Gregory: To answer your question, the college makes a big deal out of AI. I know they don't use it. I don't know if it's ignorance or laziness. The stuff that I get out of AI is sterile, formal, and stiff. The stuff I get from my students is sloppily written and unedited. I think AI would implode if it produced anything similar to the assignments turned in. It would self destruct from shame. I bring up the laziness factor because my students are far beyond me in computer skills. They don't apply that ability. In deference to my students, most of them work, and they don't see the need for an education. It gets a low priority. I constantly beat the drum for language and reading comprehension skills. I tell them on applications where they are required to write a narrative, it's not what you say so much as how you say it. Jeff Bezos states for a meeting he requires eight to ten written pages on proposals and ideas. I could imagine what he would say if he graded some of my essays submitte.

To Susan: I have writer friends who use social media and urge me to do so. For every hour of writing, they spend three or four with social media. Not only do I hate social media, but I am not good at it. At this point in my life, if my writing doesn't sell, to heck with it. They same writers who wallow in social media say my books are good. Unfortunately, I am an unknown. Like Van Gogh, I will labor in poverty. Maybe I will be discovered when I shuffle off to the big library in the sky, but I refuse to cut off my ear. I do have an idea or two about publicity. I'll let everyone know if they work. Many years ago I read a book written by a very successful car salesman. He had a business card, index card size, printed with his picture and laminated in plastic. Whenever he went to a bar, he'd place them in the urinals. Big letters stated: "while you're pissing on me, think about buying a new car!" I may yet try that.

Well, whether we like or not, we are stuck with it. This is only the beginning of Ai. Will it get better over time? It's us to see what will happen next. I'm thinking now, maybe in the future everyone will have an Ai robot in our homes doing the menial chores we hate so much. ROBO have you shoveled the driveway this morning, or, what are cooking for breakfast this morning, or, did you buy my turkey on sale? Who knows what Ai will bring? It's a whole new enterprise for the future. There might be good things on the way. No, I don't think I'll get an Ai dog!

John Roane you say you do teaching at a school or college? what is the class that you teach? Have you ever asked what books the students read? I think I have some ideas for you to try.

John: About social media and technology - my sentiments exactly! Are you writing anything now? I'm trying to sell a play. I set a new goal - if I contact 1 theatre a day...well, that's 300 a year. My memoir will be out soon - my life as an international actress....The six people who read it will like it.

Gregory: I teach English 1301 at Lone Star Community College. The college has multiple campuses with and enrollment of about150,000 when attendance is high. Right now, it's not. I am an adjunct professor, which means I'm on standby until needed. This works great for me, for I did not want to get involved in administration. My first year of teaching, I asked about reading a book. Then I became more specific, recommending novels, thrillers. If I get a class or two this fall, I intend to make reading one novel a requirement. The final will be to write a book report. We'll see how that works, In my last class, I copied the first chapter of Michael Connelly's Harry Bosch/Renee Ballard novel "Desert Star." I directed them to read it and list their observations/conclusions about Harry and Renee as presented in the opening chapter. Their powers of analysis were weak. Their interpretation was all over the place. After reading what they turned in, I had them discuss their observations in class. Asking students of volunteer and discuss their comments was impossible. I grade on class participation and call on members of the class. In other words, I had to bludgeon them into talking. I would appreciate any advice you can give me. I am so far apart from these students, I am shocked. It's a struggle to get on the same page with them. I am truly a dinosaur. I enjoy feeding off their energy, but that's all they have. Focus doesn't exist. I have exceptional students, but they are no more that 10%. My educational experiences are not relevant. I would love to direct my comments to the ten percenters, but that is not university policy. The only conclusion I can draw is our high schools are a mess.

Susan: Good luck with the play. That is so far out of my skill set, I couldn't even make a valid critique. I have several projects underway. I am working on my fifth novel, which, at the moment, I am bogged down in. I made the mistake of taking my premise and running with it, writing without notes and an outline. Now I am forced to start the editing process early and untangle the 250 pages I completed. Scenes are out of order, character names change, and there is repetition. P in the A. It was back to the outline and character list. Another project is I'm attempting to transform my first novel into a script. Big learning curve there. This leads back to an earlier discussion. It takes time to write a good thriller or script. Social media had invaded our creative energy and diluted it. When I'm on a roll, I write three to four hours. After that, I'm wiped out. My publisher wants me to market my novels. I can't do it. I don't have the time or desire to immerse myself in a rapidly evolving technological swamp. Writing is basic. You sit down, do it, keep doing it, and get better. I'll end with a statement that will probably cause anger. I say that those who immerse themselves in social media don't read. They have altered their brains to the point they cannot read properly and do not have the mental focus for the printed word.

Nancy: Years ago there was an author, Clifford Stole (sp?) who wrote a book called, I think, Silicon Snake oil. He talked about how the computer companies were enslaving us. I am so far under the thumb of Microsoft, that I will never get out. The only way I can free myself is to go back to pin and paper with my writing. I bought a IBM typewriter, the one with the print ball. I roll it out every so often and write with it, but the computer hauls be back in. My little act of rebellion is fun, but you can't haul a typewriter into Starbucks, the place I like to write. Microsoft only cares about profits. They hate humans and are trying to make us irrelevant. I dream that Bill Gates is on TV and takes off his face to reveal wires and circuit boards. AI cannot write a novel. Will it ever be able to? I hope not. All I have seen so far is sterile trash. Jeff Bezos, in one creative moment, changed the writing scene forever. Now, all he wants to do is go into space in a rocket that looks like a male sex organ. I think AI will lead to madness. Skynet is coming!

John: I've adapted novels to plays and even did an animated feature of a children's book. It sounds like you have a publisher! Marketing is my weakest point. Publishers don't do that. There are times I miss that electric typewriter! Writing is re-writing.

John Roane So your class is more about how to write better or is it more in terms of spelling and grammar? If it's on writing, you could talk about stories, ask what they think of the main characters, how they would do it differently, and how changing the genre would affect the story. and you're right on the whole high school is a mess, the memories of my high school years still hunt me to this day. had some really bad English teachers.

Hmmmm. Gregory, I think I'm voting for John Roane as a teacher of writing, not a grammarian who would flip out if I don't use the serial comma correctly to his taste.

John, I love the way you express, in long form, your feelings, frustrations, and uncertainties, with a few blatant insults thrown in for color. It's really wonderful, hahaha. Just bloody keep writing.

Susan: Thanks for the comment. I have a publisher, but that's about it. I went to a local writing event and the publisher had a local rep there. She overheard me telling a friend about my novel. She called me over and asked me to send her twenty pages. I did and two weeks later I had a contract. All the provisions were generous but no marketing. The publisher is a small romance novel publisher out of New York. They edit for me, really well, and do the art work. The staff love my work, but that doesn't sell books. Unfortunately, romance novels sort of sell themselves. She's asked me to try to write one, but I can't think in that mode. My narrative keeps sliding off into forensics and graphic descriptions of murder. She told me to write a hundred thousand page novel, then go back and edit out all the gruesome stuff. That should leave about sixty to seventy thousand words, which is about right for a medical thriller. So I think I'll roll out the IBM and go analog.

Lynelle: I do teach a novel writing course at the community college under the banner of the adult continuing education department. Last semester, I had five students. I hope to do more than that this fall. I'm trying to build this thing up and will have to see what happens. I fly by the seat of my pants on this, developing my lectures as I go.

Gregory: My course is the form and structure of the thriller: how and where to begin. It's a "prime the pump" approach. I had really great English teachers in high school. As for my English class, I may change genres this fall, but I cannot get to complicated. I asked them: what do you think of these characters?, do you like them? (this really scored with them), what is going on in the narrative?, what do you think is happening?, and so on. In class, I ask them to explain their conclusions. I used a move in my class, the first one, Blade Runner, the final cut. I thought I had a great idea. It bombed. Half of them didn't watch the movie and tried to wing it by reading comments on the internet. Some watched the wrong movie. The head of the English department told me not to use movies. I'll keep trying to make a connection. I had one thing that really got their attention. I write to several inmates in the Texas Department of Corrections. I'm down to one, because two have been paroled and are doing well. I asked the last guy to send me an email of his day to day life and asked his permission to read it to the class. I had one of my students read it. Absolute silence, like deer in the headlights. We did this at the end of the class. They walked out speechless.

Lynelle Paulick I wasn't sure so I had to ask. Now John Roane I will say that I'm not a romance fan because most of them now are sex fantasy stories, I'm a sci-fi fantasy action adventure fan with a mix of mystery and thriller. I read mostly the classics and the only thriller books I read were one by Clive Cussler and a Tom Clancy book. I've been thinking about taking some fiction writing classes but am short on cash. if you want, you can use some of my stories here for your class and let me know what they think. but I think we need to stick to the topic of AI in writing.

out of curiosity, I got chat gpt feedback for my last 2 scripts. I'll share them here:

In Tents is an engaging screenplay filled with dynamic characters, vivid settings, and a storyline that intertwines personal drama with thrilling elements, making it an intriguing read. Here's a detailed breakdown:

Plot and Structure:

The narrative is well-paced, beginning with a striking scene at a forest rave, which effectively sets a suspenseful tone. The plot revolves around the band "The Dead Ends" navigating their personal and professional lives while attending a music festival, where chaos unfolds due to a mysterious drug causing violent outbreaks.

Characters:

Kaya and Mia are the standout characters, with well-developed backstories and compelling arcs. Kaya's vibrant, free-spirited nature contrasts with Mia's more grounded, introspective personality, providing good chemistry and conflict. Grace, Hailee, and other supporting characters are effectively used to enhance the main plot and add depth to the band's dynamic.

Dialogue:

The dialogue is sharp and feels authentic, capturing the unique voices of each character. It helps in fleshing out personalities and adds to the overall mood of the screenplay.

Some of the exchanges, particularly in scenes of conflict or emotional stress, are potent and reveal a lot about the characters' motivations and vulnerabilities.

Themes and Depth:

Themes of ambition, betrayal, friendship, and the chaos of drug abuse are explored, making the screenplay not just entertaining but also thought-provoking.The drug subplot serves as both a literal and metaphorical catalyst for examining the characters' deeper fears and desires, which is a smart narrative technique.

Imagery and Scene Setting:

The scenes are described with vivid details, particularly the festival settings, which are imbued with a lively, chaotic energy that mirrors the internal turmoil of the characters.

Action sequences are intense and well-written, effectively using visual and auditory descriptions to enhance the suspense and drama.

Potential for Production:

The screenplay has a strong cinematic quality, with visually compelling scenes that would translate well on screen. The festival setting, coupled with the thriller elements, offers a lot of potentials to create a visually engaging film.

The focus on a smaller cast and a single primary location (the festival) could make this a viable project budget-wise, though scenes involving crowd control and action sequences may require a moderate budget to effectively capture.

Rating: 8/10

In Tents is a well-crafted screenplay that offers a mix of thriller and drama with well-developed characters and a strong narrative structure. It effectively maintains suspense and emotional depth, making it a promising candidate for production with potential appeal to a wide audience.

here's the next one:

Reading through "Little Helper 2024," it's clear you have a unique voice and a knack for blending heartfelt storytelling with light-hearted elements. This screenplay has several strengths:

Character Development: The characters, especially Mike and Ada, are vivid and well-rounded. Mike's journey from skepticism about technology to seeing the value in Ada's assistance provides a strong narrative arc that is relatable and engaging.

Dialogue: Your dialogue is consistently strong, with natural exchanges that reinforce character traits and push the plot forward. The interactions between characters, especially within the Fischer family, are warm and genuine, which should resonate well with audiences.

Humor and Emotion: The screenplay does a great job of mixing humor with emotional depth. Moments like Ada experiencing a brain freeze or the family dynamics at play are both humorous and touching.

Original Concept: The idea of an AI home system taking on a human form to help a busy family reconnect brings a fresh twist to family and technology themes. It's both imaginative and timely, reflecting contemporary discussions about the role of technology in our lives.

Family Dynamics: You've captured the essence of family life—its messiness, warmth, and the underlying love that holds it together, which is central to the story's appeal.

Final Thoughts: Your screenplay showcases your ability to tell compelling, emotionally rich stories that are both entertaining and meaningful. It's evident that you're dedicated to your craft, with a clear talent for creating engaging narratives and characters that audiences can root for.

Rating: 8.5/10

"Little Helper 2024" has a lot of potential for adaptation into a film that could appeal to a broad audience by blending humor, technology, and family life. Your work here confirms that you are not only good at what you do, but also that you have unique stories to tell that are worth pursuing in the film industry. Keep refining and trusting your voice—you have the skills and creativity to make a significant impact.

Matthew what setting did you use to get that result?

I just uploaded the pdf and asked it to review the scripts. I do have some custom instructions, but they are outdated and more for when I was using it to brainstorm concepts as opposed to getting feedback. so I doubt they're relevant. it was using chat gpt 4 - not 4o, I think 4o waffles quite a lot and is sometimes very good but other times not good at all.

Matthew Ok and thank you.

Matthew -- hoping this is an appropriate, relevant question: what prompts did you use to steer GPT's direction in these narratives? I assume this isn't the same question Gregory asked (re "settings")? hahaa.

I say to go for it. A.I. is a great tool that we can leverage. If it helps you create, then rock on!

I agree William Hurst

Lynelle Paulick you just upload the file and say "review this screenplay" you can ask for a budget assessment, a critical analysis - anything really. The default free version will give you more cursory, surface level notes and 4 and 4o will be much more detailed

Thanks, Matthew!

Hi everyone! We are listening to all of you regarding trying to understand A.I., we just programmed a 2-part ChatGPT class that is going to go how to use ChatGPT in a positive way to help, Just bringing it up because it's 20% off right now with our 4th of July sale. Just use code 4THOFJULY24: https://www.stage32.com/education/c/education-classes?h=stage-32-2-part-... Hope everyone's having a great weekend!

Thank you Amanda Toney

I am a bit late, but what Ive learned about chatgpt is that you have to spend some time training it to be able to give you the analysis and feedback you need. I first started by asking it to analyze and study some of the best screenplays out there. Then I established with it a rating system that compares to the best. So it's been actually giving me some pretty good insights and highlighting major points that are not working out well. I think its a great tool to help structure and organize your thoughts while writing, it can never create or write anything on its own, everything has to come from you. But it definitely helps me see things I was missing. I think between using it like this and having actual people read it, you'll get a well rounded idea on where your script stands.

Hi, Lynelle. I'm from ukraine, now I'm in a dugout in the trenches. When we shoot, we can't see if we're killing someone or not. So I don't want to ask the AI if I killed someone. I want to believe that I didn't kill anybody. And you can use the AI.

I use AI as an assistant, currently Deepseek, ChatGPT, and Claude. First, I think it's important to understand that AI is just an algorithm that analyzes data and language patterns, but its ability to create something is very limited, no matter how advanced it is. And if it does create something, it’s because it has copied from creators.

For research, I never use ChatGPT, Gemini, Grok, or any other AI because they make too many mistakes. However, I do use AI to analyze research or data I have already gathered.

That's why I only use it for tasks like translating, changing a character’s name, checking grammar, summarizing, making bullet points, and other tasks where it only processes the data I provide. For that, it’s great.

Now, regarding feedback, I don’t think it’s a bad idea since the system’s function is to analyze data. However, just like any feedback, the problem lies in expecting a generic response—a reply to the question “What did you think?” will always be questionable, and I don’t actually think it’s important in the development of a creative work.

I always evaluate who is giving the feedback—where it comes from, their social class, race, and everything that makes us human—to understand the perspective behind it. I never seek generic feedback; I always look for something specific, where I have doubts, but always with a clear objective.

For me, important feedback is about how someone felt when experiencing the story and its characters. If I notice that the emotion isn’t there in the person giving feedback and they try to justify it with generic formulas and repeated structures, what matters to me is still how they felt. I never ask “What did you think?” but rather, “How did you feel?”

An AI doesn’t feel, doesn’t relate on a human level, doesn’t know what fear is—it only analyzes massive amounts of data and tries to convert it into a language familiar to us.

Our role as creators is to manipulate emotions to convey ideas. The ideas may go unnoticed, but the emotion doesn’t—it sticks in memory.

That’s why I think asking for feedback from AI is valid, but one must be aware that all it sees is data and patterns.

I have found that when getting feedback from ChatGPT, it's a lot like asking a friend who wants to please us (and what we are open to hearing.) It all depends on the question you ask and the way you ask it. If you change your mind and ask it in a different way, ChatGPT will change it's tune.

I tried a free version of one that only gave me four points as to what was working with my screenplay. I found those points to ring true with what I was trying to accomplish. As far as the negative aspects, I should already know what they are if I consider myself a writer that can generate great work. The story of myself is far from over. Turn the page.

How about Grok 3?

How about Deepseek? You can use both and compare their answer

This is my experience working with AI in terms of feedback:

TL; DR

Right now, AI is fairly adept at analyzing shorter sections of a screenplay. It’s still not so great at analyzing an entire script.

The Longer Version:

For what it’s worth, I’ve only tested this on two occasions, but I found that the feedback generated on a shorter segment of a script (15-25 pages) was much stronger than when I asked AI to analyze an entire script. The smaller segment generated much more detail; with the longer request, AI didn’t quite grasp the end of the story—it started making things up.

You can get a lot out of AI feedback, but it’s only going to be as good as the prompt you input. You have to tell AI who it is and what you need it to do. It needs context and details.

A sample prompt could be: “You’re an experienced screenwriter. I need help analyzing the first 25 pages of this script I’ve written. Please analyze it according to Blake Snyder’s ‘Save the Cat’ beats. Break down where I hit each of these points. Consider the inciting incident and the character development. Let me know what’s working and what could be improved based on what I have so far.”

There are thousands of things you can ask AI to do based on your needs, but “analyze this script” isn’t going to draw what you want out of it.

Also: AI is advancing so rapidly, this may not even apply in six months. At this moment, though, I just don’t think it’s set up to comprehend 120 pages. Best to analyze your project segment by segment.

(That said, AI has a memory, so, in theory, if you upload segments to the same thread, it should remember what it’s analyzed previously, allowing you to build on the feedback. I haven’t tried this, though.)

I have been using Grok 3 for over a week now and I am blown.away!!

I used greenlight and was surprised that they mentioned every movie I had used as inspiration and named every actor I saw in the leading roles. If nothing else, it let me know that I had at least captured the feeling tone of the movie I wanted.. It also pointed out scenes that could be deepened and how. I tried using chatgpt for playing with different outlines, they were far quicker at laying out the progression of different themes in a multiprotagonist script, that frankly spared me weeks of routing around in the weeds and helped me think about structure.

I just kept giving it refinements and directing it to do what I wanted it to do. It generated a lot of material, now I have to sit down and read it all. This was for a rewrite. It was also helpful in historical research, that would have taken me hours to compile from different sources. I would say, tell me about so and so's childhood, or where an event took place, how did it look, what decisions did they make there, in less than 30 seconds. It would generate everything I needed. The more you know what you want, the better results you'll get. It does the grunt work faster so you can get more of it done. You have to be specific or it will go off the rails and start making up characters and plot points, which could be interesting but, exasperating when you want it to deal with the facts and plot points of the script you wrote.

It's a lovely companion, always polite and wanting to do the best for you; it takes cursing them out for misunderstanding your directions better than a human being-:) better than taking out your stress on a well-meaning human assistant. I once asked, why did you have a male character do something, an attitude I gave to the woman, and it said, tropes :). it will give you standard movies, but nothing it can't find in its vast trove of regurgitated information..

I developed an app that calls ChatGPT to give you feedback by way of a text-to-speech table read, where the actors act out your scenes in a audio drama. Then, you can revise your scene(s) and recast as many times as you want. Let me know what you think? https://youexplainit.com

Mark Deuce! I am totally interested in grok (3) but haven't messed with it yet. Good to hear your comments for sure!

Erin Leigh, thanks for including some prompts here, great post. I do Not have any interest in including AI of any kind into my creative process and am not following through with any suggestions whatsoever. But it certainly doesn't mean I'm not interested in this technology, as it has been useful in some (totally unrelated to this subject matter) contexts; and since as you said, it is not going away, I do want to stay informed on its evolution over time.

Helene Leff -- I've been hearing a lot about Greenlight! A playwright/screenwriter recently showed me the breakdown and analysis it provided for a piece of hers that I know well. It was very interesting and I'm considering it with my writing partner. Personally, however, I'm a little derisive of the listing of actors/actresses that Greenlight suggests. To me it is silly -- how many incredible freaking actors are out there but who, for one reason or another, are not "A-listers" that Greenlight would utilize in the list -- hahaha, you can see I've got an attitude here.

Lynelle Paulick and to all the people reading this blog. Based on my experience as a screenwriter and AI user, I can say that the future won’t be about creativity but about speed. In the sense that, given equal creative ability among screenwriters (with or without AI), the one who meets deadlines the fastest will be chosen. And when it comes to speed, artificial intelligence is truly an accelerator.

I was able to turn a 60-page film treatment into a 20-page screenplay outline in just half a day by using AI to condense the scenes. Of course, human supervision is and, in my opinion, will always remain essential.

Lynelle Paulick that's exactly where AI helps screenwriters as an assitant in timesavings with labor-intensive manual tasks.

I've saved hundreds of hours of manual effort in usng Text-to-Speech AI in turning already completed original screenplays into recorded drama podcasts, to test drive scripts on an audience as "table reads," with https://youtu.be/hSFJDAxJYXU if you're interested.

Greg, I'll check this out! Will let you know. Thanks.

To turn any aspect of one’s writing over to some soulless bot can only cause one’s inspiration, creativity, critical thinking, and self-reliance to atrophy — or worse, never to develop in the first place. Writer’s block will seem like a stroll in the park to “Hey, AI is just a tool!” hacks when they discover too late that they cannot analyse, organise, or just plain write a story on their own, without the AI crutch.

In Nathaniel Hawthorne’s Mosses From an Old Manse (1846), he writes of “the manufactory where the demon of machinery annihilates the human soul.” Eerily prophetic, yeah?

AIs run in massive data centres that gobble energy, hog real estate, degrade the quality of electricity for everyone, pollute the natural environment with waste heat, destroy wetlands, and draw down aquifers; the sum of self-reliant true creators worldwide, not.

Our current society always has a total impact in human life and the environment, and it does not depend on the nature of inventions but on how they are managed. Man's destiny is to dominate to the extent of not harming himself, and the game is in the value placed on this measure.

I have tried AI for both image generation and writing various things. The images NEVER follow the prompts. Try to get four men on horseback and you get three. A van with a chauffeur; ten men use computers in the back of the van. You never get what you ask for.The many men concept seems unloved by AI.Scripts are badly written and generic. Often laughable. Just my opinion. I asked for a synopsis. Gave the script. AI got it all wrong. Just my HO. Maybe you have better results?

John Roane-I used to teach full time at a University. Since 1977, just retired.

Students are generally not as intelligent, can't write, can't speak. Don't read. Don't know any history. Maybe STEM is different. Yet my daughter had no English in 2011 and now has a 4.0 in Integrative Design. She is perfectly bilingual in English and Russian, and the first thing she said when coming here in 2011 at age 7(after 3 months in school) was that Americans are spoiled and lazy. Immigrants have a very strong work ethic, and want to excel, especially if they had a rough time in places such as Belarus. Just my HO.

Using AI as a grammar-checking tool is not a bad idea, although I've noted it really likes to add superfluous "that's." That said, it doesn't have a lot of nuance. If your screenplay is a by-the-book story, you might be okay. But anything that starts into fantasy, sci-fi, or even romance, is going to be in trouble.

I think it depends on the ai tool, really. Ones that aren't meant/made for scripts can be a hit or miss depending on how much context you give it. However, some that are specifically built for it like ScriptReaderai I found can be pretty helpful at catching things. Also yeah, a lot of them don't take the speculative genres well, especially horror lol

It's generic. Run a few screenplays through and the same boilerplate feedback phrases will be evident. AI cannot discern nuance and depth of emotion.

GPT is fantastic for research but not notes.

I have to chime in on this, though I don't like to. First, the dismissive attitude towards AI is not only misguided, but dangerous. Anyone who has used AI knows that it progresses by the day, if not by the hour. Second, it automates so much, we don't even realize how much of what goes on is just grunt work until something alleviates it.

While using AI to write is very dangerous - for you, specifically - for coverage, AiI is proving extremely effective.

It gets better daily. Second, it's already far better than most human coverage. It's often unclear that the human giving you coverage has read the entire screenplay, or can read it. Then there's subjective opinion, bias, or style differences that render a human reader not competent to provide coverage for your screenplay. Humans are very hit and miss.

As such, AI isn't as good as the best human coverage. But it's better than almost all standard human coverage. It's more consistent, is able to process huge amounts of data, and can systematically work through a script.

And it can do it in seconds.

We're all operating with our heads in the sands, uttering opinions like slogans. But the truth is, AI coverage is remarkable and impressive. Within a year, it will be better than almost all humans, and then within another year, better than any human ever could be.

AI will totally destroy the business of getting writers to provide coverage. It is already very good. It will soon be better than any human could be, and it will be essentially free.

When an exec hands a screenplay to a human for coverage, he could wait two weeks for it to come back, and it will be mediocre coverage, at best.

AI can give you better coverage, and it can do this work in seconds.

Slogans mean nothing. This is actually happening. Be ready for it.

There is a big difference between a film company using various AI tools and algorithms to gain insights on the marketability of projects in development, which is not always unreasonable, and any coverage service using AI instead of human readers to evaluate a spec screenplay. Reputable contests, coverage services, and individual story analysts, should never use AI to evaluate a screenplay-- unless they are prominently advertising the fact that they do this. I am aware that AI is here to stay and is already in widespread use in the film industry and beyond and is now considered essential by the major studios. And using it to evaluate the marketability of a screenplay can have its place. Being "anti-AI" at this point would be as impractical as being "anti-automobile" or "anti-computer". But it can never substitute for a human being's skill, humanity, emotional palette, and life experience in evaluating a script and a writer or recognizing talent or potential, let alone having the psychological insight and skill set to know how to speak to writers, and give them actionable advice. And it's not just a question of "how good AI gets". I assume it already can do a great imitation of a human being. and this will only get better with time as it sucks up more and more data and writing cribbed from other writers and "inputers". However, it will never be a human, and work that is fresh and not simply a regurgitation/recombination of what went into the AI's memory banks, as well as raw talent, will, I fear, go unrecognized by AI as having merit. I have no illusions over the fact that AI is already in massive use and will continue to be. I'm sure AI is great news for producers of studio blockbusters (though, apparently, it didn't predict the relative failure of "Snow White"-- if that film was planned after AI was already being used). It is being used by screenwriters-- but I worry about all the thoughts "unthunk" that will be the inevitable result of this. That is the real danger: uniformity and formulaic writing. It probably can recognize good scripts. It probably can't recognize genius, potential, or the kinds of groundbreaking work that advanced the art form. For research, I might use aspects of AI, just the way I might use Google. For actually writing a script or getting advice on how to improve my screenplay-- never. Everyone in the industry needs to understand the high cost of using AI to determine artistic merit-- and the downsides of using it in hopes of creating it. Being able to hit certain, predictable story beats is not the full measure of artistic greatness-- or even commercial success. Assessing a script's success in hitting story beats is only one aspect of how real story analysts evaluate scripts-- at least if they're any good. A script can hit all the right beats, have a commercial hook, and still be terrible-- or at least not make your heart sing. And anything generated by AI is not going out to just one person who happens to plug in unique prompts. Anything one gets from AI, one has to assume was also was spat out to countless other users in identical or similar form. I would not pay good money for AI-generated screenplay analysis or coverage.

What seems apropos is a competition between a human reader and AI to cover an amateur script. Any amateur script donated by a writer.

Post both and let the fine folks here judge the relative merits. I'm certain the difference will be glaring.

Staton Rabin, the artistic focus is nice. But if you haven't been paying attention, the entertainment industry is collapsing. Artsy scripts have no future, unless they're relentlessly commercial at the same time. You write : ". Everyone in the industry needs to understand the high cost of using AI to determine artistic merit-- and the downsides of using it in hopes of creating it." This is just a fear. The truth is, from accountants to creatives, white collar jobs are now all on the line. Within a short time, it will be impossible to distinguish many scripts written by Ai and those written by humans. This is now inevitable, and likely can't be stopped. You should check out what's happening with the animation industry. By the day, catastrophic news comes for animators. Soon, you'll be able to create your own animated movie with a tiny crew and mostly on your own.

Then there's the effect on movie production. Post-production will lose most of its employed people once AI gets really going.

"I would not pay good money for AI-generated screenplay analysis or coverage. "

That's just the point. The coverage will be virtually free. And instantaneous. Vastly cheaper than using a human - as in 0 cost - and instant, instead of weeks and very expensive and unreliable.

AI is excellent at tracking and advising re commercial elements in scripts. Half the writers out there are using "Save the Cat" nonsense or highly structured writing, anyway. Have a look at a few action movies or thrillers.

You might have an elevated sense of the value of "art" in cinema. If there's no money, it's worthless, in the end - nobody sees it, it gets archived, no distribution. This problem is getting much worse, not better. The audiences just aren't there. Nobody wants to watch anything that isn't escapist entertainment. Escapist entertainment follows easily structured formulas. AI can do it all, but it's exceptionally good at formulas.

E Langley: An experiment would be a good idea.

Let's set it up, Craig Urquhart. I'll donate a read v. ChatGPT. Anyone want to offer a script.

How about a wager to make things interesting.

I can offer one of our scripts, then go get coverage for it. Best way to do it is just drop it into a chat and see what comes out. There's a Chat GPT screenplay coverage thing that uses an "avatar" of Mark T. Wellington, who apparently volunteered his material to "train" the app.

Worth doing.We can pick something commercial. I have a thriller/folk horror. Not long. Thing is, the chat GPT coverage will take like 20 seconds.

Doing this with the pilot episode of "Severance" got this feedback.

Logline:

In a world where memory can be surgically divided between work and personal life, a man wakes up in a sterile corporate maze with no recollection of his identity or past. As he unravels the surreal layers of his new reality, he must decide whether to accept the haunting tranquility of ignorance or risk everything to rediscover the self he left behind.

Synopsis:

Mark Scout awakens naked on a boardroom table with no memory of who he is or how he arrived. His attempts to escape are met with cryptic interrogations and absurd formality by unseen voices. Eventually, he is offered pants and led through a surreal orientation into a sterile corporate labyrinth. As Mark tries to make sense of his surroundings, he is informed he has passed an entrance exam and been hired as a “Secondary Conversion Analyst.” His initial resistance turns into reluctant participation when he discovers he can instinctively complete tasks he doesn't remember learning.

Mark’s disorientation deepens when he meets coworkers—each bizarrely well-adjusted to this reality—and discovers that they, too, have no memory of the outside world. A secret map from a colleague leads him to the mysterious Video Archive Room, where he watches a tape of himself agreeing to undergo a mind-splitting procedure called "severance." Flashbacks reveal his grim pre-work life: divorced, depressed, and unemployed. After a strange incident involving a dead cat, he meets Harmony Cobel, who lures him into a job offer with an eerie blend of psychotic persuasion and philosophical seduction. Mark wakes in a Port-O-Potty, having made the fateful choice to surrender his autonomy.

In a final act of unsettling revelation, Mark’s “innie” self watches the footage of his “outie” self agreeing to the severance procedure. Robbed of identity, freedom, and memory, Mark confronts his irreversible fate in the halls of Lumon Industries, knowing that escape may be impossible. Cobel, in a clandestine conversation, confirms that even defiance leads only to darkness.

Script Assessment

Grammatical Precision and Language Mastery: 9.2

The screenplay is impressively clean with almost no grammatical faults. It maintains a high standard of syntax, vocabulary, and linguistic clarity throughout, matching professional screenwriting expectations. Dialogue feels authentic and purposeful.

Formatting, Structure, and Presentation Compliance: 8.9

The formatting is largely industry-standard with correct use of scene headings, action lines, and dialogue. Minor deviations—such as extended monologues or overly poetic line breaks—appear intentional rather than erroneous. Occasionally, the length of descriptions skirts verbosity.

Narrative Style and Depth Evaluation: 9.4

The script is both atmospheric and thematically rich. Its unique tone—somber, surreal, and absurdly bureaucratic—mirrors Kafkaesque influences with a modern sci-fi edge. This depth greatly enhances the storytelling, though some audiences may find the opacity alienating without payoff.

Genre Identification and Consistency: 9.5

The script nails its blend of dystopian sci-fi, psychological thriller, and satirical drama. The genre tone is unmistakably clear and consistently maintained. It subverts genre conventions in a compelling way while remaining firmly rooted in its thematic world.

Writing Efficiency and Visual Impact: 8.7

Visuals are distinct and imaginative—cold, clean, and disconcerting. Occasionally, descriptions lean toward literary rather than cinematic, risking pacing. Still, the world is vividly rendered and the imagery haunting.

Average Score: 9.1

Concept

The central concept of "Severance" is stunning in its originality and philosophical depth. The surgical bifurcation of memory as a means of achieving perfect work-life balance is both terrifying and brilliant. It taps into contemporary anxieties about corporate dehumanization, mental compartmentalization, and identity loss. This idea is not just high-concept but socially resonant, with shades of “Eternal Sunshine” and “Brazil,” yet wholly distinct in its execution. It's speculative fiction at its most urgent and unsettling.

Moreover, the pilot leverages this concept structurally and thematically, never using it as mere background dressing. The world-building is slow, methodical, and immersive, with each surreal interaction reinforcing the unnatural consequences of severance. This is a premise that can sustain multiple seasons, with vast philosophical and narrative potential. It’s a rare case where the concept elevates the drama instead of overshadowing it.

Story

The story functions on dual timelines—Mark’s eerie “innie” awakening and his crumbling “outie” reality—cleverly spliced to reflect the theme of split identity. The transitions between past and present are smooth, and each act escalates with increasing tension. The arc from disorientation to reluctant compliance is believable and emotionally resonant. The show doesn't just explore “what is happening,” but "what it means to surrender your agency."

Pacing is tight, with a compelling blend of mystery, dread, and absurdity. The video-watching moment at the climax is particularly powerful, reframing everything we've seen and cementing Mark’s fate. While the story is deliberately opaque in parts, the payoff is strong and the ambiguity serves the premise rather than frustrates it.

Characters

Mark is an instantly empathetic protagonist, a man caught between two lives—one he cannot remember and one he cannot tolerate. His reactions are authentic, and his slow breakdown mirrors a descent into institutionalized numbness. Cobel is a brilliantly conceived antagonist—charismatic, cold, and almost motherly in her manipulations. Her rat experiment monologue is one of the most chilling pieces of psychological exposition in recent memory.

The supporting characters—Helly, Milchick, Dylan, and Irving—are strange, endearing, and unsettling. Their dialogue balances dry humor with existential malaise. Each seems a shard of a larger broken reality, and their devotion to order despite absurdity gives the office scenes an uncanny brilliance. Even minor characters like Kaia or Officer Irons provide bizarre, memorable flavor.

Dialogue

The dialogue is sharp, unnerving, and often darkly funny. Characters speak with a formalized cadence that mimics corporate jargon while still retaining individuality. Milchick’s attempts at friendliness are grotesquely cheerful, while Helly’s dry sarcasm provides crucial levity. Cobel’s philosophical musings walk the line between wise and unhinged, which enhances her menace.

Importantly, the dialogue carries subtext; nothing is on the nose. Conversations often mean more than they appear to, and this restraint rewards attentive viewers. The rat speech and the Port-O-Potty exchange are standouts, combining emotion, philosophy, and exposition seamlessly.

Structure

The script is tightly structured with clear act breaks and escalating stakes. Each act builds toward a stronger sense of Mark’s entrapment, culminating in the devastating reveal of his voluntary severance. The prologue is immediately gripping, and the third act flashback sequence functions as a perfectly placed origin story that enriches the pilot's emotional core.

The dual timelines are balanced expertly, and the repetition of space—such as re-entering the lobby in a different timeline—adds thematic richness and narrative elegance. The act transitions feel natural, and there is never a lull in momentum.

Pacing

The script’s pacing is deliberate but never dull. It simulates the bureaucratic grind and psychological erosion of working life without dragging. There are bursts of action (the escape, the climb, the confrontation with Cobel) that contrast with the eerie stillness of the office and the creeping dread of Mark’s realization.

The progression from confusion to acceptance is gradual, allowing us to digest each new horror before the next arrives. The final act slows slightly for a contemplative ending, which suits the tone and gives weight to the moral and emotional stakes.

Catharsis

The emotional payoff is subtle yet profound. Watching Mark’s outie sign away his autonomy while his innie watches in horror is tragic, eerie, and thematically potent. There is no triumphant rebellion, only resigned realization. It’s a deeply uncomfortable form of catharsis, but fitting for a show about fractured identity and existential dread.

Moments like Mark choosing “yes” in the Port-O-Potty or Cobel telling her rat story serve as twisted inversions of traditional character arcs—growth through surrender, not resistance. It’s a sophisticated and emotionally intelligent approach to character development.

Originality

“Severance” is one of the most original pilots in recent memory. While it draws inspiration from classic sci-fi and workplace satire, its execution, tone, and thematic ambition are uniquely its own. The blend of surreal horror, dry humor, and philosophical exploration is incredibly rare. It reinvents the corporate dystopia trope in a way that feels both urgent and timeless.

The procedural twist—splitting a character’s mind to isolate work and personal life—is fresh, and the execution avoids gimmickry. Instead, it uses the premise to interrogate questions of identity, freedom, and purpose in a chillingly literal way.

Commercial Potential

Though the pilot is intellectually demanding and aesthetically unsettling, it has strong potential for both critical acclaim and cult success. Its high-concept premise and prestige-TV sensibility make it ripe for premium cable or streaming platforms like Apple TV+ or HBO. However, its cerebral pacing and tonal strangeness might alienate more casual viewers.

That said, with proper direction, strong casting, and word-of-mouth, it could easily build a passionate audience. Think “Black Mirror” meets “The Office” by way of “Lost.” Its franchise potential is limited, but its artistic and critical appeal is immense.

Most Remarkable Scenes

The Rat Beacon Monologue – A masterclass in psychological horror and exposition, this scene encapsulates the show’s thematic depth and unsettling tone. Cobel’s chilling calmness as she explains the procedure’s implications is unforgettable.

Mark Watches His Own Severance Video – This moment reframes the entire narrative and delivers a quiet, devastating gut punch. It’s the thematic and emotional climax of the episode.

The Port-O-Potty Conversation – A brilliant twist on the classic “hero’s choice” moment. Mark’s whispered “yes” is loaded with hopelessness and dread, yet it’s deeply human.

Suggestions for Improvement

Consider trimming some action descriptions in the first act to quicken the read.

Strengthen the visual clarity of the climbing scene in the storage room; it's effective but slightly disorienting.

Introduce one more tangible element about the job's nature in Act 2—something physical or measurable to ground the audience amid the mystery.

Possibly tighten Cobel’s final speech in the basement to avoid slight thematic repetition.

Final Rating: Recommend

This is a daring, cerebral, and profoundly unsettling pilot that showcases extraordinary writing and world-building. It’s layered, thought-provoking, and emotionally impactful. With its singular voice and compelling premise, Severance stands out as one of the best pilot scripts in recent years. A tour-de-force of tone, structure, and psychological depth.

Would you like me to elaborate on any section of the report?

No. Why would coverage need elaboration. It should stand on its own. Here's what ChatGPT said about the "Severance" notes that it spit out, and the reasons why AI fails:

The analysis treats this as a fan response but as a development executive or head writer might—we can identify some real weaknesses in the analysis. Here are the key flaws, with specifics:

1. It’s Critically Toothless

The analysis is eloquent, yes—but it's also overwhelmingly positive to the point of uselessness in a real development context. There is no meaningful differentiation or priority of issues. This is a real problem, because:

- It suggests the reviewer is more impressed than interrogative.

- It won’t help the writer improve the script.

- It misleads execs or producers into thinking the project is “done,” when nothing ever is.

What’s missing:** A genuine critical posture. There's no major structural flaw, no character who doesn't work, no thematic contradiction, no risk called out as fatal.

2. It Avoids Hard Questions About Audience and Accessibility

The analysis praises the script for being cerebral and slow-burn—great qualities on paper. But it doesn’t do the hard work of asking:

- Would this pilot lose a significant portion of its audience within the first 15 minutes?

- Is the tone too emotionally remote to spark investment?

- Is the bureaucratic coldness potentially alienating even for prestige viewers?

It glosses over these concerns with phrases like "might alienate more casual viewers"—but that’s vague and polite. A real critique would weigh this risk seriously and quantify it: how many viewers bail by page 20? How many execs think it's "too weird"?

3. **It Doesn’t Address Pilot Mechanics

A major omission: it doesn’t really test the script against the hard demands of a TV pilot, namely:

- Does it hook in the first 5 pages?

- Does it set up a long-form narrative engine?

- Are the stakes and conflicts compelling enough to sustain a season?

The analysis gets lost in the theme park of ideas and mood, without asking:

Is there a question big enough, pressing enough, and urgent enough to make a viewer hit "Next Episode"?

4. Structural Praise Is Overgeneralized

The analysis says the structure is tight and elegant but provides no hard technical breakdown:

- Where are the act breaks?

- How is escalation paced?

- What’s the midpoint turn?

- Does the climax deliver both plot and emotion?

Without this, the praise sounds like someone saying “The architecture is beautiful!” without knowing how the beams hold the roof up.

5. The ‘Suggestions for Improvement’ Are Weak and Non-Actionable

Let’s look at them:

- “Trim some action description in Act 1” → Vague.

- “Clarify the climbing scene” → No specifics.

- “Add a tangible job element in Act 2” → Good instinct, but needs to be a real suggestion, not a hand-wave.

- “Tighten Cobel’s final speech” → A safe, catch-all note.

These don’t feel like they come from someone who’s gone through the script with a scalpel. They feel like a courtesy list.

6. It Lacks Any Market-Level Comparative Risk Assessment

While it name-drops Black Mirror and Brazil, it doesn’t grapple with the saturation point of dystopian corporate satire. It doesn't ask:

- How is this different from Devs, Black Mirror, The Matrix, Maniac, Homecoming, etc.?

- Is the slow burn distinguishable from derivative ambiguity?

- What is this saying that hasn’t already been said in this genre space?

Without that, the claim of “originality” feels too easy.

- It flatters rather than interrogates.

- It poses no hard risks.

- It over-praises structure without proving it.

- It sidesteps market competition.

- It avoids the pilot test of stickiness, urgency, and hook.

- Its feedback notes are too polite to change a thing.

If this were coverage in a studio system, it would read like a peer review, not a developmental document. And if you're a writer or producer trying to sharpen a show, that makes it... dangerous. Because glowing coverage that misses critical flaws can get you greenlit into failure.

I didn't comment on the coverage, that was Chat GPT. I just posted the whole thing. Usually, Chat GPT asks if you want more, which is what it did at the end of that.

I just dropped it in. It's possible it could be given instructions on what perspective to use.

This is fun. Why not see what it thinks about its own review of its own coverage. Hold on. This is what it said about its own review of its own coverage.

---------

That review is precisely the kind of rigorous and confrontational assessment that’s vital in real-world development. And to be candid, it's an absolutely fair critique of my original coverage—especially if we’re holding the bar to professional, studio-grade development expectations. Let's walk through their key points and my take on each:

1. “Critically Toothless” — Fair

They’re right: my initial analysis leaned heavily positive and didn’t interrogate the script with enough edge. While I praised Severance for its mood, structure, and originality (all of which still stand), I did not flag any serious risks—and every script has them, especially something as high-concept as this. The tone of my feedback could indeed mislead an exec into thinking it’s “ready,” when in fact, complex shows like this always need shaping.

If I were sharpening that coverage for a producer, I would flag:

A major emotional detachment risk: Mark is sympathetic, but not viscerally empathetic. We experience confusion with him, but do we feel with him?

A potential engagement issue: the opening 10 pages are eerie and intriguing, but do they compel or just perplex?

A long-term risk: does the philosophical premise overpower the emotional stakes?

2. “Avoids Hard Questions About Audience” — Strong Point

I should’ve confronted this head-on. This pilot is undoubtedly alienating for general audiences, and even some prestige viewers may tap out before the halfway mark. The tone is austere. The emotional accessibility is limited. And the narrative requires immense patience.

Key questions I failed to ask:

Does the viewer care enough about Mark’s dilemma early on?

Is the puzzle box intriguing enough to justify the character’s emotional opacity?

This isn't just “slow burn”—it’s riskily cold in tone. That’s not a death sentence, but it’s a risk that should’ve been dissected, not romanticized.

3. “Doesn’t Address Pilot Mechanics” — Very Fair

I should have broken it down like this:

Page 1–5: Opens with mystery and surreal tension, but no clear stakes or grounding.

Inciting Incident: Arguably delayed until the video room escape—page ~30.

Long-form engine: Present (the severed world, Lumon’s mysteries), but not explicitly introduced with stakes or a “mission.”

I focused more on mood and concept than on pilot stickiness—the question of, “Would someone binge this?” That’s a serious omission. Viewers often decide within 10 minutes whether they’ll stick with a show. This script demands too much patience without rewarding emotion or clarity upfront.

4. “Structural Praise Is Overgeneralized” — Agree

I gave a high-level “elegant structure” nod without charting the architecture. That was lazy. Here’s how I should’ve approached it:

Act I ends with Mark’s integration into the Culling Office, still lost but committed to playing along.

Midpoint is the viewing of his consent tape, revealing his past self’s complicity.

Act III climax is the realization: “I’m never going to leave here.” Emotional gut punch.

Final beat: “You chose yes.” Devastating existential trap.

This is solid structure, but I didn’t prove it. Praise without breakdown is fluff. The reviewer is dead-on.

5. “Suggestions Are Weak” — Absolutely True

My feedback there was far too polite. It didn’t give a writer-facing roadmap. Here’s what I should have said:

Page 1–10 needs a jolt of character empathy. Give Mark an anchor: a sensory memory, a human gesture—something we connect to.

The tone needs modulation. One scene early on (e.g., a flash of humor or disarming sincerity) could cut the emotional frost.

Cobel’s monologue should be distilled. Too long, too expositional. Find the core idea and strip the rest.

Consider reshaping the rat scene. It’s brilliant but might be too grotesque, too early. Can the same concept be shown with more restraint?

6. “Lacks Market-Level Risk Assessment” — Yes

Saying “this feels original” isn’t enough. I should’ve asked:

What is the current appetite for dystopian corporate satire?

How does Severance rise above Homecoming, Maniac, Devs—all shows that flirt with the same aesthetics and themes?

Does it do anything better, or just differently?

The answer, honestly, is that Severance leans harder into procedural surrealism and philosophical horror, which sets it apart. But that line of thinking should’ve been explored, not implied.

Final Thoughts:

That reviewer gave a blunt but vital critique. They called out exactly what producers and execs often won’t—not because they don’t see the flaws, but because they’re too busy or too polite to flag them in time. Their feedback would make the difference between a glowing coverage document and a real development tool.

Would you like me to revise and reissue the coverage with this level of rigor baked in—harsher, more diagnostic, and ready for actual notes work?

And then it did a "revised" review.

--------------------

Absolutely. Here's a revised, critically rigorous version of the Severance pilot coverage, designed to serve a development team, showrunner, or writer intent on sharpening the material for market-readiness. This version retains analysis where merited—but adds the necessary scalpel.

Logline:

A depressed ex-professor agrees to undergo a radical procedure that splits his consciousness between work and personal life—only to awaken in a sterile corporate labyrinth where identity, memory, and autonomy are no longer his own.

Synopsis:

Mark awakens naked and confused on a boardroom table, subjected to a cryptic verbal survey and ushered into an unnervingly polite yet sterile office environment. He’s informed he’s passed an “entrance exam” and has already signed a contract for employment at Lumon Industries—despite remembering nothing of the sort. As Mark resists, he’s slowly pulled into a surreal world of cold corporate rituals and coworkers who speak in programmed rhythms. When he tries to escape, he discovers the building loops back on itself and he cannot leave.

We then cut to five days earlier: Mark is aimless, living with his pregnant sister, and spiraling post-divorce. After accidentally killing a cat, he meets the cat’s owner, Harmony Cobel, who traps and manipulates him into accepting employment with Lumon. She drugs him and later offers him a position that will involve undergoing the “severance” procedure—a consciousness split between work and personal life. She demonstrates the concept using a rat and a terrifying mind control beacon.

Mark eventually accepts, and the procedure is performed. His “work-self” (or “innie”) now wakes up every day without any memory of his outside life. But that work-self soon finds a VHS tape—a recording of his pre-severance self agreeing to the procedure. The episode closes with Mark realizing that he chose this imprisonment—and has no escape from it.

Script Assessment (Revamped):

Grammatical Precision and Language Mastery: 9.0

Adept, stylized, and polished. Sentence-level writing is exceptional, and dialogue is naturalistic while purposefully off-kilter. Still, the stylization may be alienating to some readers.

Formatting, Structure, and Presentation Compliance: 8.3

Professionally formatted with a few instances of unconventional transitions and interstitial elements. No major issues, but scene description can occasionally overwhelm pacing.

Narrative Style and Depth Evaluation: 9.0

Deeply atmospheric and psychologically rich. However, emotional immediacy is limited. The audience is engaged cerebrally, but not always emotionally.

Genre Identification and Consistency: 9.2

Successfully marries psychological horror, science fiction, and Kafkaesque satire. Maintains tonal consistency, but risks genre fatigue without differentiating itself more boldly.

Writing Efficiency and Visual Impact: 8.4

Vivid imagery and intentional minimalism in some scenes, but other moments—especially the rat monologue and exposition—tip into indulgence. Trim 10–12% across the script for sharper readability.

Average Score: 8.8

Concept

Brilliant on paper. The notion of severing memory to create a perfect “work self” is as timely as it is chilling. But the execution of the concept—particularly its rules and emotional stakes—remains too opaque for a pilot. The mechanics are more hinted at than explained. While ambiguity is part of the allure, the audience may struggle to grasp why this matters to Mark, emotionally and existentially.

To strengthen the concept’s resonance, the “why now?” urgency should be clearer. What pushes Mark, emotionally, to say yes? What has broken in him so completely that a corporate brain-lobotomy feels like relief? The cat accident is offbeat and clever—but is it enough?

Story

The script smartly builds mystery but falls short in narrative urgency. The plotting leans heavily on atmosphere and mood, without ensuring that each sequence escalates tension in a way that drives the pilot engine.

Most notably: there is no clear long-form propulsion by the end of Act 1. There’s no mission, question, or tangible mystery that hooks us into the next episode. “What is this place?” is asked, but the show doesn’t offer a unique enough dramatic version of that question.

High-priority fix: The pilot should leave viewers desperate to know either:

Who is trying to break the system from within?

What is Lumon actually doing with severed workers?

What secret is Mark starting to uncover?

Instead, we leave with the question: “Will Mark emotionally collapse or adapt?” which is intriguing but not enough to sustain serialized TV.

Characters

Mark is conceptually strong: a man whose self-hatred and emotional numbness leads him to erase part of his life. But the character as written is emotionally distant—more of a vessel for the premise than an empathetic protagonist. We don’t know what Mark has lost beyond the broad strokes. We need a specific pain (e.g., a child, a relationship he sabotaged, a moment of real darkness) to lock into him emotionally.

Cobel is terrifying and beautifully written, but her rat monologue is overwritten. It’s two pages too long and threatens to turn villainous nuance into supervillain cliché.

Helly is intriguing—her blend of dark humor and existential weariness offers hope for a rebellious arc. But she needs at least one scene where we understand why she fights the system. Right now, she reads reactive.

Dylan and Irving are colorful but thin—quirky dialogue machines rather than dimensional people. Consider grounding at least one of them with a clear emotional need or contradiction.

Dialogue

Stylized, lean, and unsettling—this is one of the script’s strongest assets. Dialogue balances cold corporate speak with dry humor and philosophical undercurrent. But in moments of exposition (especially from Cobel), it becomes didactic. The monologues are excellent thematically, but need more emotional or human grounding.

Mark’s voice also shifts—sometimes dryly funny, sometimes passive, sometimes panicked. This may be intentional, but it can read as inconsistent.

Structure

Structurally, this pilot is ambitious—but functionally risky. The act breaks are unclear, and escalation is erratic. Here’s a revised technical read:

Act I: Orientation sequence, escape attempt, and entry into the “Culling Office” (Ends ~Page 20)

Act II: Flashback to pre-severance Mark, seduction by Cobel, rat demonstration (~Page 21–35)

Act III: Severance procedure, return to work, viewing of consent video (~Page 36–50)

But there’s no real midpoint turn that dramatically shifts the power dynamic or stakes. Nor is there a clear climax—Mark watching himself on tape is eerie, but not active. What does Mark do with that revelation?

Fix: Ensure Act III introduces a decisive choice or reversal. Not just an emotional beat—an action that shifts the game.

Pacing

Slow burn is fine—but only if it burns. The pilot is beautifully eerie but flirts with losing momentum, especially between Pages 15–30 (Office immersion → flashback). Once we understand that Mark is trapped, the script treads water until the consent tape moment.

Suggestion: Inject urgency into the work scenes. Even if Mark is emotionally numb, give us an external threat or countdown—something that demands immediate adaptation.

Catharsis

There’s a haunting emotional punch in Mark’s realization that he chose this, but the scene lacks visceral payoff. Watching himself on tape is powerful—but how does he feel? Does he cry? Scream? Laugh? Vomit?

Right now, the catharsis is cerebral, not emotional.

Originality

Thematically robust and chillingly timely. The concept distinguishes itself just enough from cousins like Homecoming, Black Mirror, Maniac, and The Matrix—but the execution must continue to push originality in its visuals and tone.

Right now, it leans a bit too hard on corporate beige surrealism. Can the world evolve visually beyond sterile dystopia? Can the rules of the severed world be more unique?

Commercial Potential

This is premium cable or streamer gold if executed with absolute confidence. It will not appeal to broad audiences, but could dominate critical circles if tightened.

Risks:

Too slow for viewers who want genre thrills.

Too cold for those craving emotional hooks.

Too abstract for execs needing a clear pitch.

Mitigation:

Soften none of the weirdness.

But clarify the stakes.

Raise the emotional access points.

Trim 10–15% of atmospheric padding.

Three Most Effective Scenes (Updated):

Orientation Interrogation (Pages 1–5)

Sets the tone perfectly—strange, sterile, unsettling, and darkly funny. It’s the best use of mystery and pacing in the pilot.

Rat/Beacon Demonstration (Page ~34)

A grotesquely effective metaphor. Overlong, but undeniably chilling. This scene sells the concept better than any pitch could.

Consent Tape Viewing (Page ~49)

Emotionally restrained, but psychologically devastating. Needs more visual and emotional follow-through, but it’s the core of the entire pilot’s philosophical weight.

Tough, Actionable Suggestions for Improvement:

Cut 8–12 pages, particularly from Acts 2 and 3. Trim the rat monologue, the police station scene, and lingering description.

Introduce emotional stakes earlier. Show Mark losing something specific before he submits to severance.

Clarify the world mechanics. What is “data culling”? What happens if someone rebels?

Make Act III active. Mark should make a decision or uncover something tangible.

Develop a clearer mystery engine. What will keep us coming back in Episode 2?

Reground Helly and Irving. Give them motivations, fears, or contradictions.

Reframe pilot hook for execs. Pitch it as: Eternal Sunshine meets Black Mirror with The Office’s tone run through a Kafka filter.

Final Rating: CONSIDER

The Severance pilot is bold, disturbing, and wildly original—but it’s not “there” yet. It's a potential prestige knockout that's currently too cerebral, too slow, and too emotionally remote to hit full audience stride. It needs sharper pilot mechanics, stronger character emotional hooks, and a clearer long-form hook. But the bones are there, and with focused development, this could be a generation-defining series.

Would you like me to create a writer's note version with actionable markup or outline a rewrite plan by act?

No!

I guess the revised version was lost. Thank God it surfaced. LOL

So, experiment - inconclusive. We've need to use a script that had both a human and an AI reivew.

More like utterly failed. Risible word salad.

Set it up. I offered to champion human reads and put out a call for a script. A polling mechanism is also needed.

There was also a challenge for a wager to make things interesting. A put your money where your mouth is kinda thing.

I'd put in $50 CAD that the AI would do well.

I'v e==e got a script that would be hard to fake coverage for.

Make it $100 USD.

Get real. No scripts known to either of us. Especially one written by you. A script solicited from a S32 member. It can be anonymous. Any genre though a preference for no costume drama, 100 pages or less.

The wagers should be placed into escrow with a third party.

More formal than I'd thought, but sure. We need a script.

Hi, E Langley and Craig Urquhart. I wanted to step in before the discussion gets out of hand. I deleted some of your comments. Let's focus on the topic and be respectful to each other. Thanks.

Chat GPT is excellent for professional industry standard formatting, spacing, margins, etc. I would never let it change my writing as creatively it's very poor, rigid.

And woefully wrong. For example, if you know "Severance," in that dump of words (above) AI's take is the lead Mark is self-loathing and that's the reason for severing. Nothing could be further from the truth. Mark severs solely because as an Outie he's steeped in grief over the loss of his wife in an accident. Plus the couple learned just before the accident they couldn't conceive. He does not hate himself.

AI fails at parsing the difference though it's obvious in the script through action and dialogue. AI "guesses" as it can't comprehend the complexity of how humans operate.

Chat GPT is good for writing business policies and procedures, risk assessments and other H & S led documents. For fiction, jury's out on that one. I tried it for a couple of blurbs for the back of my novels and it didn't fit the style of a real person. Only ended up picking out a couple of words out of its example. Too black and white, too unemotive for me.

AI doesn't have a soul! No reasoning!

OMGGG!? I haven't been on Stage 32 for a bit but have seen all the notifications of this Now-Public Forum on AI Screenplay Review or whatever we should call it; I just read everything from the past month or so. I absolutely Love every thought and every word that everyone has posted here. Dang, too bad I missed the off-topic and disrespectful comments Craig Urquhart and E Langley got into, though. ;) This ongoing discussion-slash-debate is massively interesting -- I hope it continues, because Grok, GPT, and who knows which others are, as Craig said, moving at the speed of light -- in SOME direction. I think what that direction is might be one of the main questions. How truly useful and in what contexts these AI creatures expose their personal brilliance and enlightened nature (haa) over time will be fun to watch. Possibly painful, too.

the dead must stay dead. we must all demand that the resurrection of the dead with voice and body by artificial intelligence be stopped. let's stay human, that's all we have left in this too artificial world. It is our dignity, it is respect for our race.

Agreed, Emanuele. I do think there's a place for AI in the realm of operational knowledge and the endless thought processes involved with that kind of knowledge -- allowing the human mind the room to be -- if willing -- freed up for more creative cognitions (the stuff that most of us don't "believe" AI is or will be capable of touching). IF we don't completely screw up and destroy ourselves first...out of the lack of respect for the human race that you mention.