We’re documenting the decisions, mistakes, costs, tools, workflows and very human work behind turning an original story into a finished series using traditional and AI filmmaking tools.
Why?
There are now approximately eleven billion videos online showing what the newest AI video generator can do:
- Beautiful woman walks through neon rain.
- Spaceship flies over alien city.
- Viking stares meaningfully into the middle distance.
Lovely. But that’s not the question we’re interested in.
We want to know something much harder:
Can you actually build a professional original series using AI as part of the filmmaking process?
Not a 10-second clip. Not a fake trailer for a movie that doesn’t exist. Not a collection of gorgeous shots held together by dramatic music and optimism.
A real series.
That also means characters who remain recognizable from scene to scene. A visual language. Performances. Story continuity. Editing. Music. Sound design. Production deadlines. A budget. Copyright considerations.
And, inconveniently for many AI generator people, a story people actually want to watch. With a story you can lean in to, a plot that holds your interest, and characters you can root for, hate, love and long to watch in the next episode to see what happens.
So we’re doing it. A laboratory.
Our test case is an original series called “Deadly Flowers”, and we’re building it in public.
“WHY DEADLY FLOWERS?”
Deadly Flowers is a stylized murder mystery built around four women, buried betrayals, disappearing lovers and a town where memory itself may not be trustworthy
Visually, it gives us plenty to wrestle with: recurring characters, specific costumes, multiple locations, atmosphere, mystery, symbolism, flowers, murder and enough continuity problems to make any AI filmmaker quietly reconsider accounting.
Which makes it perfect.
If these tools are going to become part of professional filmmaking, they have to survive more than just images transformed into 8 or 10 second videos. They have to make a meaningful difference as a tool for storytellers to get closer to a “yes” and get greenlit.
“AI ISN’T THE PROCESS. IT’S A PART OF THE PROCESS.”
One of the strangest things about the current AI filmmaking conversation is how often people think the software is the star. They seem to forget that without a story to make the images worth watching, it’s just a bunch of data that becomes a bore.
We’re much more interested in what happens after the generation button gets pushed.
A generated image is not a finished shot. A generated video clip is not a scene. A collection of clips is definitely not a film. It all starts with story. Your story.
Then our work happens in the decisions.
- Which character design works?
- Which one gets rejected?
- Does the performance fit the scene?
- Does the camera angle support the story?
- Does the wardrobe stay consistent?
- Does the location suddenly change architecture between shots?
- Does the actress appear to have quietly become someone else?
- Can we fix the shot?
- Do we need to edit the image and regenerate?
- Should we regenerate it or start with a new prompt?
- Should we abandon it completely and solve the scene another way?
That’s filmmaking. A thousand decisions to create something memorable. Magical. New. The tools may be different. The decisions haven’t changed. The biggest difference is that one person may be wearing assorted hats including “written by”, “set design”, “costume design”, “cinematographer”, “sound designer”, “sfx designer”, “director”, “ producer”… you get the idea.
“THE SCREENPLAY STILL RUNS THE SHOW”
This is the part we believe writers have a major advantage in. AI can generate images. It can generate movement. It can generate voices, environments, music and visual ideas at extraordinary speed.
What it cannot do for you is decide what your story means.
That starts in your brain and what you write on the page.
Every scene still needs a purpose. Every character needs a want. Every episode needs movement. Every shot should support the story rather than exist because the software happened to make something pretty.
That sounds obvious until you watch a lot of AI films. Pretty or explosions or flying or sexy women, for the sake of sexy women, can become a trap.
A beautiful shot that does nothing for the story is still a useless shot.
“WHAT WE’RE DOCUMENTING”
Throughout production, we’re going to show the parts people usually skip over. We’ll share what works, but the failures are probably going to be more useful. We’ll look at things like:
- How we create and maintain consistent characters.
- How many generations it actually takes to get usable material.
- What we regenerate versus what we repair manually.
- How we direct AI-generated performances.
- How we build scenes rather than isolated shots.
- How much the production actually costs.
- Which tools save us time and which ones quietly eat entire afternoons.
- How editing changes generated material.
- How we handle sound, music and voices.
- How we document human creative contribution for copyright purposes.
- What we would do completely differently next time.
Because “Look what AI can do” is rapidly becoming the least interesting part of this conversation. We want to know what filmmakers can do with it.
“BUILDING THE HUMAN RECORD”
We are also documenting the human work performed throughout the project.
At Greenlit Studio, that includes story interpretation, visual development, character development, selection and curation, compositing, manual modification, editing, sequencing, sound design, typography, revisions and final assembly.
We call this our “Human Creative Contribution Record”.
That matters creatively, but it also matters legally. The U.S. Copyright Office continues to distinguish between AI-generated material and human-authored creative expression. Human modifications, creative arrangements and other sufficiently expressive human contributions may be protectable even when AI-generated elements are included in the finished work. (U.S. Copyright Office).
That means the question isn’t simply, “Did you use AI?” The real question for a copyright “yes” is, “What did the human creator actually create?”
We think filmmakers should be able to answer that. And we always want to protect our clients.
“WHY BUILD THIS IN PUBLIC?”
Because there are plenty of predictions about where AI filmmaking is going. Not our focus.
We’d rather know where it genuinely reduces production barriers, where it creates new problems, where human craft becomes more important, and where the current technology simply isn’t ready yet. We also want writers watching us to see something larger than a software demonstration.
For most of filmmaking history, getting a story from screenplay to screen required money, access, equipment, crews and permission. Those barriers haven’t disappeared.
But some of them have moved.
Why that matters.
A writer can now begin proving what a story looks like before someone else agrees to finance the entire thing. And make it easy to visualize because it’s actually visually accessible.
We’re not trying to replace Hollywood. Just change the opening move. And we know from experience, that’s worth paying attention to.
So we’re building Deadly Flowers. We’ll show you what works. We’ll show you what breaks. And occasionally, we’ll show you the thing that looked absolutely magnificent right up until the character grew a surprise third hand. (It recently just happened in one of our generations – even now with much better generators…akh!)
Welcome to the Greenlit Studio production diary.