Can AI and Human Storytelling Coexist in Filmmaking? I Put It to the Test
A New Tool in a Long History of Innovation
Filmmaking has always evolved alongside technology and with every leap forward, the same questions emerge.
What happens to the artists? What does this mean for the future of the industry?
This has always been a business shaped by change. Especially by technologies that democratize the tools of creation. From the invention of film itself to sound, to television, to digital editing, each advancement has disrupted the status quo while expanding the artform’s reach.
It all started with Edison and the kinetoscope, introducing motion pictures to the world. Then came sound. In 1927, The Jazz Singer debuted as the first motion picture with synchronized dialogue using the Vitaphone. While some celebrated it as a breakthrough, others resisted worried it would ruin the purity of silent cinema. As captured in the book The Endless End of Cinema:
“Talking motion pictures will never prevail.”
—Jack Warner, qtd. in Warner, 1926“(…) things will be back to normal, and they [distributors] will be considering the quality in a silent product instead of going crazy over a lot of useless and irritating noises.”
—Walt Disney, qtd. in Gabler, 2007
(as cited in The Endless End of Cinema, Sergi & Rydstrom, 2020)
Sound didn’t destroy film—it elevated it, unlocking deeper emotional immersion.
Then came television, shifting how we consumed stories. Later, digital cameras and nonlinear editing blew open the doors for independent filmmakers, drastically lowering costs and putting cinematic tools into more hands than ever before.
Now, we’re standing at the edge of the next evolution:
AI in filmmaking.
Why I Said Yes to Making a Fully AI-Generated Scripted Scene
When friend and colleague Eliav Mintz (co-founder of Little Island) asked if I wanted to create a fully AI-generated scripted scene, I said yes without hesitation.
Lately, my feed has been overrun with fear-driven headlines and hot takes:
“AI will kill Hollywood.”
“Pixar’s in trouble.”
“The movie industry is finished.”
As a working filmmaker and editor in docs and scripted, I don’t believe that narrative. But I won’t lie to you and say I’m not scared when I see these narratives come across my feed. Like, what IF AI replaces artists?! And I’m not alone.
Like hundreds of other creators I know, I’m not here to celebrate the end of artists. I’m here to explore how AI tools for video editing, VFX, animation, and storytelling might support our work, not erase us from its creation.
Humans have always been at the center of storytelling—on cave walls, through music, in paintings, photography, and now, even AI. But tech doesn’t feel. It doesn’t grieve, fall in love, or experience longing. It can generate images, but it doesn’t live life. That’s where humans come in.
Too much of the AI content flooding social media feels hollow. Its pretty imagery without story, emotion, or rhythm. So I wanted to find out: What happens when a human filmmaker uses AI, but still approaches the process with intention, pacing, and feeling?
Editing Is Innovation at Its Core
Filmmaking and technology have always been intertwined. My own craft—editing—is built on innovation.
Editing is the transformation of disconnected images into meaning. It’s where emotion is built. As Martin Scorsese put it:
“Editing is the heartbeat of the film, where raw emotions are sculpted into a resonating masterpiece.”
So Eliav and I set out to build a scene entirely using AI tools for filmmaking from visuals to voice to animation. He focused on the script and generations while I took those and cut the story together. This wasn’t an experiment as proof to remove people from creating, but to explore where the human touch is still essential in the process. Check it out below to see the finished version:
Torch image created using ChatGPT and later animated with RunwayML
The Tools We Used: A Multi-AI Workflow in Action
To create the finished scene, we used a mix of AI tools across the entire post-production pipeline:
Midjourney was our starting point, used to generate the core images and character designs.
For the dialogue, we paired Dreamina for AI lip sync with ElevenLabs, one of the leading tools for AI voice acting, to generate the spoken lines. While the voices were serviceable, they still lack the emotional nuance of human performers.
In Adobe Premiere Pro, we used Generative Extend to extend shot handles—essential for refining pacing and giving room to cut.
For a key moment foreshadowing the line “with shadows behind you,” I needed a quick insert of a torch and Hades’ shadow passing. I used still references from our previous Midjourney generations, found a visual I liked, then imported it into RunwayML to animate. It took several iterations to get the shadow pass working, but it helped shape a moment that felt intentional.
Lastly, Kling continues to improve when it comes to realistic AI camera movement and character animation, adding dimensionality to otherwise static frames.
Adobe Premiere timeline using the new Generative Fill Tool to create extra handles for clips.
What I learned is this: AI is a powerful tool—but it’s just that, a tool. Human storytellers be it actors, directors, editors, VFX artists are still absolutely necessary. But AI can help us move faster, reduce costs, and lower the barrier of entry for emerging voices. To me, this isn’t about removing the artist contrary to some trolling posts. It is about empowering them.
“What I learned from this process is that humans are still absolutely essential. These AI tools aren’t production-ready yet, but they’re incredibly useful for pre-visualization and for blocking out entire films or TV shows before anything is actually shot.
This kind of planning can drastically reduce production costs for studios that don’t want to spend extra days experimenting on set.
A creative team with strong editorial foundations can help save studios hundreds of thousands—if not millions—by avoiding footage that would otherwise be wasted, and by arriving on set with a clear, near-final vision before a single day of principal photography begins.”
— Eliav Mintz
This is a proof of concept, and there’s still room to push it further. But here’s what I personally feel about the tech so far:
What Works:
AI can create stunning visuals and animate them pretty well from a single text prompt.
Lip sync is surprisingly solid, especially for how complex it is.
Character consistency across shots is possible even without a custom-trained LoRA.
A solo artist or small team can produce an entire scene at a desk. The only limit is imagination, time, and patience.
What Still Needs Work:
AI can’t act (yet). Performances tend to be exaggerated and lack subtlety.
AI voices lack emotional range. They can sound passable, but they don’t feel real.
Two-shots and character interaction are difficult to maintain consistently across generations. a LoRA may be able to change this. We’ll see these tests soon.
Camera angles and eyelines are inconsistent, which can break continuity and immersion. Again, this I assume is being worked on now. But audiences are trained from a hundred + years of filmmaking rules. Even when working with AI, you still have to know and play by the rules audiences expect before trying to break them.
This process reminded me that while tools evolve, the heart of storytelling remains human. If we keep artists at the center, AI doesn’t have to be a threat—it can be a breakthrough. I implore all of you to go out and try interacting with some of these tools and see for yourself what is possible.