How I Made an AI Film: River of Inheritance

Author(s): Sophia Banton

Originally published on Towards AI.

How I Made an AI Film: River of Inheritance

I love films. I love Old Hollywood. I love animation — and stories with heroes.

I randomly decided to enter a film festival after animating a few of my AI-generated images. With no story or concept in mind, I sat down at my laptop with three weeks to create something. The brief was simple: produce a film that was one to ten minutes in length and incorporated AI in a meaningful way.

Initially, I felt like a one-minute film was all I could manage. To my surprise, I ended up with a 6-minute, 11-second film equipped with a full soundtrack. A testament to how far AI and access to AI have come.

Keep in mind, while I call myself an AI storyteller, I have no formal training in the art of film. I’m actually a scientist and an AI professional.

Here’s how I built River of Inheritance.

Watch the Film

Here’s a 2-minute excerpt of River of Inheritance. This is not the full festival submission, but a short version showcasing the visual tone and style of the film.

Setting

The story takes place entirely outdoors and is centered on a river. Water is the recurring visual anchor — from the calm surface of the river to stormy ocean waves. I aimed to capture both fantasy and realism, using AI-generated images to set the visual tone and animation to bring the elements to life.

Image Creation

I used Google ImageFX for image generation.

I created thousands of images by repeating, refining, and adjusting prompts until I found the ones that felt right. It was a process of iteration and intuition. In the end, I selected 31 final images to build the film.

Every image was custom prompted with no presets or shortcuts. I focused on composition, wardrobe, and tone. I wanted scenes that looked lived in, not overly stylized.

Character Design

I designed characters by prompting specific details related to wardrobe, fabric type, color palette, and environmental consistency. Each final character was the result of multiple iterations. I revised prompts until textures (like folds in fabric), skin tone realism, and lighting felt accurate and believable within the context of the environment.

One of the characters emerged as the visual anchor of the story — a woman featured repeatedly throughout the film. She became the narrative thread across scenes, appearing in both full-face and profile views.

What stood out most was how well the tools rendered and animated people of different ages — elders, adults, and children — with impressive nuance. Skin tones were represented across a wide range, and the movements of each group felt grounded and lifelike. These results highlight the progress of AI and suggest its growing potential as a viable tool for the film industry.

Animation

Once I had the images, I animated them using Runway Gen-3 Alpha.

My options were to generate either 5- or 10-second videos. I chose 10-second videos to give myself flexibility in editing. I trimmed some of them to fit pacing and transitions.

I kept motion prompts simple and clear: water, wind, waves, storms, and cloth movement. The goal was fluidity, not chaos. The water animations were designed to move realistically so the film would feel natural and human.

For human movement, I followed a similar approach — prompting specific expressions like confidence and pride, or actions such as smiling, laughing, and eating.

Editing

I used CapCut Pro to arrange and finalize the video. I added the 31 final animations from Runway and then ordered the videos based on my storyboard. For emphasis on some scenes, I slowed the speed of the video so that the character had more screen time.

Sound and Music

I layered ambient sound (rain, rivers, ocean) with instrumental tracks and natural environmental sounds, like birds. There was no dialogue — an intentional choice, because I wanted the river to “sing” as the main supporting character of the film. I found a song in CapCut’s audio section that matched the river’s tone and message. For additional scenes, I used other tracks to better capture the on-screen activity and then returned to the main song.

Tools I Used

• Google ImageFX (image generation)
• Runway Gen-3 Alpha (Turbo) (animation)
• CapCut Pro (editing/export)
• Royalty-free audio libraries (sound)

Challenges

Time constraints:

• I had just three weeks to go from idea to execution.
• There was no script or storyboard at the start.

Image selection:

• I created thousands of images.
• Narrowing down to 31 usable ones took constant testing and refinement.

Character consistency:

• Generating the same character multiple times was one of the biggest hurdles.
• I used highly specific prompts — around facial structure, hairstyle, and clothing — to create consistency.
• This allowed me to show her full face in several scenes.
• In others, I relied on her profile or rear view, combined with consistent attire, to help the viewer recognize her.

Prompt sensitivity:

• Small changes in wording could drastically change results.
• I had to carefully test and revise prompts to keep facial features, lighting, and wardrobe consistent.

Final Output

• Runtime: 6 minutes 11 seconds
• Format: 1080p
• File size: 500+ MB
• Submitted to Runway’s 2025 AI Film Festival

What I Learned

This project taught me how powerful AI can be when paired with clear creative direction. I learned to:

• Prompt with intention, knowing that small wording changes can completely shift the outcome.
• Think visually and structurally, even without a traditional film background.
• Work within limitations and still tell a cohesive story.
• Build consistency across AI-generated outputs through iteration, not shortcuts.
• Trust my instincts — even when there’s no roadmap.

Am I a Thief?

That question lingers over every project created with generative AI. I used AI to generate images and animation, but I also spent hours crafting prompts, refining visual details, building continuity, and shaping the emotional tone of the film.

So, what does authorship mean in this context?

I didn’t draw every frame by hand. But I directed every element — composition, costume, expression, movement, sound. This wasn’t automation. It was the intention.

Maybe the better question is: If I stole something, why does it look so much like me?

River of Inheritance tells a new story — a story I created about my heritage and my ancestors. It includes scenes that have never been depicted in mainstream media and characters shown with dignity rather than caricature.

In my opinion, this isn’t theft. This is what it looks like to use AI to preserve culture and educate each other about who we are and where we come from.

Conclusion

I was already dazzled by the progress and capabilities of AI image generation tools, but the animation tools have deepened my appreciation and hope for the future. What I experienced making this film — and what we are now experiencing culturally — is the beginning of a new era in film and media.

A new generation of storytellers, many without formal artistic training, will be able to tell stories we haven’t heard before, or reimagine familiar ones with fresh perspectives. With powerful, accessible tools at their fingertips, they will build new worlds, one prompt at a time.

What once required teams of hundreds can now be done with a string of prompts. Tools like these are lowering the barrier to entry — and raising the ceiling for creative experimentation. That reality is both unnerving — and empowering.

About the Author

Sophia Banton is an Associate Director and AI Solution Lead in biopharma, specializing in Responsible AI governance, workplace AI adoption, and strategic integration across IT and business functions.

With a background in bioinformatics, public health, and data science, she brings an interdisciplinary lens to AI implementation — balancing technical execution, ethical design, and business alignment in highly regulated environments. Her writing explores the real-world impact of AI beyond theory, helping organizations adopt AI responsibly and sustainably.

Connect with her on LinkedIn or explore more AI insights on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI