A 15-minute short film, fully generated with AI in @runwayml , by a Cannes and Busan-selected filmmaker.
A mother. A storm. A sea that listens.
The creative choices, the edit, the tone, the direction of acting and emotion: all human. AI is the tool, not the author.
It's already been hugely useful to us for building proofs of concept, unlocking financing, testing a narrative or a visual identity before committing the real budget.
We're doing exactly that in Cannes right now. Successfully.
Took us 8 months to get a meeting with a studio that owns major horror IP, and only after they'd read our deck and the full script.
A suite at the Ritz Carlton during EFM Berlin. Top execs in the room.
3 minutes in, the head of the studio cuts us off, stands up and says: "I'm sorry, we can't do that."
She'd just seen something in the cover image that she has "a personal phobia" of.
That was it. Meeting over, project out.
Ombak is a 15-minute short made with Seedance 2: a supernatural psychological thriller about a young Indonesian stewardess stranded at sea, rooted in local folklore.
We used it at Cannes last month as a proof of concept to move the feature forward.
Full film: https://t.co/MqOEDnPXbX
In my experience, two things get hard fast in AI video once you put several characters and layered action in a single shot: consistency, and holding the realism and quality.
So I experimented. What worked on Ombak, my Seedance 2 short: generate the background action as its own clip first, then use it as a reference for the main shot.
What worked on Ombak, my Seedance 2 short: generate the background action as its own clip first, then use it as a reference for the main shot of the lead crossing the room with that clip as reference. Seedance picked up the dialogue and adapted the pacing of the main shot to it. The speed of her walk matched the length of the conversation behind her on its own.
And since the reference clip carries the background, the prompt can stay focused on the foreground action. No overloaded prompt, no drop in quality, and the yacht's architecture stays consistent throughout the film.
Martin Scorsese is backing a new AI company and using the technology to storyboard movies.
“Cinema is a young medium, only around 125 years old, so we have to be open to how it can evolve.”
https://t.co/WSNH6uVExf
This is the model that actually produces something. A style and a process take real time on real projects, not one-week turnarounds. I built mine across a feature that played Cannes and Busan and released theatrically in 28 territories, and I’m rebuilding it now AI-native on new projects in development. That depth is exactly what we’re looking to commit to.
AI labels alone “do not affect how our videos are recommended or whether they can earn money. This is purely about giving viewers the right information at the right time.”
YouTube is rolling out a new internal system to help identify AI-generated content.
“If a creator doesn’t specify whether or not they used AI, but our systems detect significant photorealistic AI use, we will now automatically apply a label,” YouTube said.
https://t.co/sQdtSissBk
@JSFILMZ0412 Same path here. Ombak, also made with Seedance, 15 minutes, full narrative structure. Used as POC to finance the feature, screened privately at Cannes Marché du Film, financiers impressed. The long form works, industry starts responding.
https://t.co/kxkk0AFxZL
@eliott__mogenet The interesting inversion is that scarcity used to discipline taste. You couldn’t waste the take, the day, the location. Abundance removes that discipline and reveals which filmmakers had internalized it and which were merely subjected to it.
@runwayml Made Ombak entirely with Runway. 15-min proof of concept that just screened privately at Cannes Marché du Film. Financiers impressed, feature is moving forward. None of this would exist without AI. We tried for years.