Netflix-owned Eyeline Studios is using generative AI to replace locations, rebuild crowds and change lighting after filming, while developing tools that could even let directors reposition the camera once an actor’s performance has been captured.
“We’re not prompting our way into shots and movies. That’s not who we are,” Mohsen Mousavi, Eyeline’s VP of Virtual Pioneering, said at Turin’s View Conference Tuesday, describing an approach the studio calls “hybrid production,” combining live-action photography, conventional VFX and generative workflows.
Mousavi, whose credits include the final season of “Game of Thrones,” argued that producing a persuasive AI image is only the beginning. Generative models can get a production roughly “60, 70%” of the way, he suggested, but “getting to the 100% is what matters the most.” He was describing the difficult final stretch between promising output and footage that meets the consistency, resolution and creative-control demands of professional post-production, rather than a formal performance metric.
Eyeline’s approach starts before cameras roll. Its generative art department explores designs alongside traditional artists, while “sparse set” workflows allow performers to act on partial physical sets under relatively flat lighting. The surroundings and lighting can then be rebuilt, with the actors’ original performances retained. “It’s an extension of the traditional art department,” Mousavi said, rather than a replacement.
A related volumetric-capture system allows camera perspectives and lighting to be reconsidered after filming. In demonstrations, Eyeline showed footage shot in one location and transformed into another, including environments created without relying on green screens. “It gives us both freedom of light and freedom of camera after shooting performances,” Mousavi said, emphasizing the importance of preserving actors’ identities and performances.
The studio has already applied parts of the approach to Netflix productions. On “Brazil ’70: The Third Star,” the drama about the national team’s 1970 World Cup victory, Eyeline helped build the vast, emotionally responsive stadium crowds required to recreate the tournament. The process combined filmed elements, sprites, conventional CG and generative techniques, rather than depending on AI alone. Mousavi said the results “would have been very, very difficult to do within that budget” through traditional VFX methods.
For Netflix’s five-part historical documentary “The American Experiment,” he demonstrated background extensions, generative dust and other digitally constructed elements integrated with practical footage. Eyeline also showed techniques for changing a performer’s apparent age while retaining the underlying acting, as well as a research short exploring post-shoot camera and lighting adjustments.
Among the technologies still being developed are controllable generative fluids. Current generative simulations are unreliable, Mousavi acknowledged. “They’re not deterministic, and they’re very hard to control,” he said. Eyeline wants artists to adjust water’s velocity, viscosity and turbulence with keyframes, and add foam or spray without altering the underlying simulation. “Often in visual effects, what you really want to do is that you want to break the physics of the fluid,” he explained, to achieve a specific creative effect.
Other demonstrations included restyling a Gaussian-splat 3D capture from winter into autumn and AI-assisted relighting of human performances. One system uses 40 synchronized RED V-Raptor cameras to collect training data; Eyeline has also developed more portable capture equipment. Mousavi also showcased research into constructing a navigable 4D scene from a single-camera recording. Such experiments, he stressed, are distinct from techniques already deployed on completed shows.
Eyeline’s systems rely heavily on proprietary work rather than off-the-shelf generators. “Most of the work we do is based on our own research and our own models,” he highlighted. The studio also uses outside partnerships and specially captured training data, though Mousavi did not identify its foundation-model providers.
The presentation comes as Netflix separately expands its AI ambitions, including its acquisition of Ben Affleck’s filmmaking technology company InterPositive. The streamer said in July that roughly 300 titles had already employed some generative AI.
For Mousavi, technical capability alone is insufficient. “We want to make sure that it’s not a black box,” he added, insisting that cinematographers and production designers must be able to shape the results. High dynamic range, 4K finishing and temporal consistency remain significant challenges, even as Eyeline advances its own solutions. “The goal is to make it completely cohesive,” he concluded, regardless of whether a final image originated in camera, traditional VFX or generative technology.