Netflix has used artificial intelligence (AI) in some form on about 300 of its titles, Lee Sung-gyun, the streamer’s senior director of APAC production, told the Busan International Film Festival (BIFF), per variety.com.
On the same panel, Korean director Kang Yoon-sung said AI workflows could reduce a feature film’s production budget by 30% to 40%.
Lee spoke alongside Kang, director of “The Outlaws,” and Kimball Thurston, chief technology officer at Wētā FX, at a session titled “Film Production and AI: New Tools, New Possibilities”. Netflix and the Motion Picture Association (MPA) co-hosted the event, part of Creative Asia, the Netflix and BIFF programme for emerging Asian talent. Film critic Yun Sung-eun moderated.
Lee said Netflix uses AI in previsualisation, production design, set-dressing concepts, editorial and visual effects (VFX). He said new technology has helped with “how” films are made but never decided “why” a story is told, and that AI is no different.
He screened a reel from Eyeline, Netflix’s VFX studio, on what the company calls hybrid production, which combines live-action photography, VFX and generative tools. A second clip showed AI and VFX producing large crowd scenes for the Netflix India title “Glory”. Lee said such scenes would normally have been cut because of tight budgets and schedules.
Kang co-directed “Run to the West” with AI filmmaker Kwon Han-seul. The film opened in Korean cinemas last October and was billed as the country’s first feature to use AI across its entire production. He said the technology was rudimentary then and was used mainly for compositing and creatures, which did not blend well with the live footage. In a year, he said, its quality has reached the point where commercial use is no problem.
Kang’s test clip for “The Gulf of Aden”, a film about military operations at sea, was made entirely with AI this year using the Chinese software Kling. He said it was rendered at reduced quality, adding that current tools can output 4K at 10-bit colour or higher, which is sufficient for cinemas.
Asked about cost, Kang said calculations made for fundraising on “The Gulf of Aden” showed a reduction of around 40%. He said the saving came not only from replacing CG and VFX, but also from a simpler shoot and a shorter production period.
He said AI has revived projects that collapsed at the budget stage, opening the door to genre and sci-fi films at a time when the Korean industry is struggling to grow. However, he warned that AI often removes the discussion, effort and experience of many people on which filmmaking has relied, leaving images based on one person’s ideas. “In terms of delivering emotion and moving audiences, I think we’re in a period where it has actually gotten worse,” he said.
Lee said he strongly agreed, adding that “a human has to touch it once more” to convey emotion.
Thurston described the deep-learning denoiser behind his second Academy Scientific and Technical Award as “assistive AI” that “touches pixels, but it only fixes them” and cannot alter creative intent. He said Wētā has used machine learning for at least a decade, or 20 to 25 years counting early versions of Massive, the system built for “The Lord of the Rings”. He also flagged gaps in colour accuracy, saying models trained on video images do not reproduce the correct red of a Coca-Cola can.
On keeping pace with new models, Lee said Netflix works under strict internal guidelines and tests tools before adopting them, adding that the company is “not wildly obsessed” with speed. Thurston said Wētā is building an open-source tool through the Academy Software Foundation to track AI use and provenance, designed to link with initiatives such as C2PA.
On sustainability, Thurston said Wētā’s data centre in Wellington, New Zealand, which uses closed-loop cooling, now draws “zero liters of water per year”. Lee said Netflix considers sustainability company-wide, including its 2030 carbon-reduction goal, and does not study AI’s impact separately.
Kang expects AI use to split in two directions: short-form content will focus on productivity, with some Chinese projects already made entirely with AI, while film and drama will focus on credibility. He also showed footage from a feature animation he is developing, with photoreal backgrounds and animated characters he wants to move like live-action actors.
On replacing actors, Kang said it would be difficult because actors are creators who deliver emotion. He sees AI’s role in stunt and stand-in work instead. Thurston agreed, saying: “The point is not to replace actors, right? No way.”
Lee closed by saying technology will keep changing and surprising audiences, but “it’s humans who will write”, adding that Netflix wants more stories told by humans.