Generative AI can widen the options for a visual effects shot, but artists still decide which of them serve the story, according to James Rustad, global head of GenFX at Eyeline, Netflix’s in-house VFX studio. Rustad made the case at an InnoAsia session “Making What’s Next: Generative AI and the Future of Filmmaking at Eyeline,” held at the Busan International Film Festival‘s Asian Contents & Film Market. He used Netflix’s “Brazil ‘70: The Third Star” as a reference point.
The series features more than 1,300 visual effects shots, including over 800 set on the soccer field. Eyeline works with 45 artists on a hybrid approach, Rustad said, bringing generative tools into a process that also draws on established effects techniques. Asked which aspects of filmmaking AI cannot replace, Rustad pointed to how quickly the technology is advancing, and steered the discussion toward where generative tools belong in filmmaking today and which tasks they should handle.
His presentation mapped four stages of a generative AI shot workflow: setup, narrative, look and delivery. Prompts, reference gathering and initial tests establish the options available. Those results are then assessed against the story before moving through visual refinement and final delivery.
Lighting, scene details and period accuracy remain part of that refinement, followed by compositing, upscaling, bit-depth checks and technical review. The workflow places a narrative decision between initial experimentation and visual finishing, moving from “what’s possible” to “what works for the story.” That step is where artists’ judgment comes in: selecting which generated results should become finished shots.
Extract — continue reading at the source.
