Testing out AI filmmaking

This week, I carved out some time to dive into a few AI filmmaking tools—and put two of them to the test: ArtList and LTX.

The process itself is refreshingly simple: write a descriptive prompt, let the AI work its magic, and within moments, you’ve got a generated video clip. The more detail you pack into your prompt, the more the AI has to work with—which makes these tools incredibly useful for building mood trailers or rapid shot lists to pitch ideas to creative teams.

But there’s a catch. The deeper you go into refinement—tweaking prompts, adjusting scenes, chasing that perfect shot—the faster your token credits evaporate. That sting is especially sharp when you’re working with a free tier.

Original prompt for LTX – A young white European hipster girl walks down a street in Fort Greene, Brooklyn. It’s nighttime, as she leaves a trendy bar.

The seven‑second clip is pretty solid. It doesn’t really scream Brooklyn—more like any NYC block—but the AI character actually looks okay.

In my mind, i wanted the girl to a little more blond and less east villagie. So my next prompt was – A young white European hipster girl walks down a street in Fort Greene. She is blond with a sun dress on.

As you can see, the shot has completely changed. This time, she’s walking on a different street—one that feels more like Union Square—but we no longer see her leaving a bar. The vibe is solid, and the medium-shot format remains consistent.

Next, I tried a different AI tool—and this time, I gave the prompt more action-oriented details. That may have been the issue, because the output shifted from a single continuous shot to a rapid-fire sequence in the style of Michael Bay: quick cuts, but with far less nuance.

Here’s the prompt for this shot –
A young white european hipster girl walks down the street in Fort Greene Brooklyn.  It is nighttime and she is leaving a trendy bar.   She has blond hair and a nice sun dress on.   She walks down block to the subway and heads down to the platform.  As she gets to the turnstile she realizes she misplaced her wallet.  She takes out her cell phone but the phone battery dies as she tries to pay with it.

So the overall vibe feels rushed—especially at just four seconds. I do like the reaction at the end, but as you can see, that single prompt burned through the rest of my credits. On the previous example, I ran out right after adding the hair color and dress update.

For a first attempt and creative inspiration, AI filmmaking is great for generating solid reference videos. But if you’re aiming to produce an actual short film, I’m fairly certain you’ll need a lot of time to refine your vision—and even more tokens to get anywhere close to it. Next time, I’ll try creating a dialogue scene and see how that plays out. In the meantime, I’ll have to wait until my free account replenishes before I can do that.

Latest

Down
DOWN the motion picture.
April 20, 2026

Down

Down is a hypnotic psychological drama exploring queer isolation, addiction, and the search for meaning in this post pandemic world, where real connections are hard to find. Set in the borough of Brooklyn, DOWN follows Glenn Solemon, a middle-aged man navigating the blurred lines of desire, addiction, and power. Haunted by a past lover and […]
Read More
Basics International project
February 25, 2021

Basics International project

This past winter, I spend some time in Ghana and got to meet some great people. One of my highlights was meeting , Patricia Wilkins, who runs this non-profit organization that helps kids get an education. Here’s a video I did for the organization.
Read More
Proactive Content
February 23, 2021

Proactive Content

I had the great pleasure of running Proactive’s NYC studio for 2.5 years. In that time, I help create, produce and manage all the videos productions, from conferences to in-sudio and zoom videos. Below is a selection of past works, broken down by type. The video below were created in-studio, utilizing Tricaster’s TC1 with three […]
Read More