Notes from My Workshop at Spikes Asia: The Experience Gap in AI Filmmaking
A massive “Thank you!” to Melanie Speet (Director of Spikes Asia) and Ed Pank (SVP Cannes Lions) for their trust and support in bringing this workshop to life.
Three Questions About the Future of Filmmaking
Last week at Spikes Asia, I had the pleasure of running a workshop titled “From Script to Frame.”
The session was essentially a guided tour of how I now build motion scenes using generative tools — the everyday working process that has quietly become part of my practice at RockPaperScissors, where I build Story Engines for Brands.
Near the top of the workshop, I asked a question that stopped me for a moment.
“How many of you have ever been on a real film set?”
Only about 20% of the room raised their hands. Almost all of them were older. Most clearly over thirty. Then I asked another question.
“How many of you have made AI video clips before?”
About half the room raised their hands. Then a third question.
“How many of you have made a video longer than one minute?”
Not a single hand.
That moment told me something important about where we are right now with AI filmmaking.
We are entering an era where many people learning AI film craft have never experienced film craft itself. That changes the way you think about what’s happening in this moment with generative video.
Because filmmaking has always been a craft learned through proximity. Which leads to a question that feels increasingly relevant in the age of AI.
What does filmmaking look like when the tools arrive before the experience is gained and the craft is developed?
Learning Filmmaking Without Film Sets
For most of the history of filmmaking, learning happened in a very physical way. You learned by being there. That’s how I learned it.
Standing on set. Watching the lighting crew shape a scene. Seeing how directors talk to actors. Watching camera operators adjust framing by millimeters. Sitting through painfully long pre-production meetings where everything — every prop, every angle, every movement — gets debated and planned.
Even if you weren’t directing, you absorbed the language. You learned what coverage means. You learned why blocking matters. You learned how many decisions sit behind even the simplest shot.
You learned the rhythm and rigour of production. It’s why, when I sit down today to generate a scene with AI tools, I’m not thinking only in terms of prompts.
I’m thinking in shots. I’m thinking about coverage. About how a camera moves through a scene. About where actors enter the frame. About what the lighting probably looks like outside the frame.
All of that thinking comes from years of simply being around real production.
And it turns out that experience translates surprisingly well into AI.
AI Is Powerful — But It Still Needs Craft
One of the big myths right now is that AI filmmaking replaces traditional filmmaking knowledge.
In my experience, the opposite is happening. The more you understand how films are actually made, the better you become at directing AI tools.
When you know how a dolly move works, you can describe it better. When you understand coverage, you stop generating isolated clips and start building scenes. When you understand blocking, you stop thinking in terms of “images” and start thinking in terms of action within a frame. AI can generate visuals incredibly quickly. But it doesn’t automatically generate storytelling.
Storytelling still requires intent.
The One-Minute Barrier
The other signal from the workshop was equally interesting. Many people in the room had experimented with AI video tools. But their work stopped at short clips.
Eight seconds. Twenty seconds. Thirty seconds. No one had built something longer than a minute. Which makes sense, i guess. The tools themselves encourage experimentation at that scale. But filmmaking really begins when you move from clips to scenes to sequences, when shots start connecting, when pacing starts to matter, and sound design begins shaping the emotional rhythm.
That’s when prompting stops being experimentation and becomes storytelling.
A Suggestion for AI Filmmakers
If you’re interested in AI filmmaking, here’s a piece of advice I shared in the workshop.
Spend time on a real film set. You don’t need to be directing. Just observe.
Watch how long it takes to light a scene. Watch how many people it takes to move a camera. Watch how carefully directors plan even the smallest actions.
Sit in on a pre-production meeting if you can. Those meetings are where filmmaking actually begins. Long before anyone presses record. Because what you’ll discover is that filmmaking has never really been about cameras.
It’s about decisions. AI doesn’t remove those decisions. If anything, it multiplies them.
From Script to Frame
That’s ultimately what the workshop at Spikes Asia was about. Not just how to generate images, but how to think like a filmmaker. To move from idea to script to scene to frame. To story.
Because the tools are evolving incredibly fast, but the fundamentals of storytelling — framing, rhythm, motion, performance — remain surprisingly constant.
In the end, the technology is new. But the craft still matters. And the better we understand that craft, the more powerful these tools become.
RockPaperScissors - Story Engines for Brands
I build Story Engines for brands — systems that turn strategy into narrative, and narrative into films, campaigns, and editorial platforms.
Increasingly, those engines combine traditional creative craft with AI-driven production workflows, allowing ideas to move from script to screen faster than ever before.
If your organization is exploring what storytelling looks like in the age of AI, let’s talk.