AI Film Costs: Token Bills Threaten to Outpace 35mm Budgets

Studio executives in South Korea warned on November 8 that artificial intelligence might soon make feature films more expensive to produce than they were during the 35mm era. Speaking at the Taiwan Creative Content Fest in Taipei, industry insiders detailed how the mechanics of token usage create financial risks that traditional shooting does not face.

Kim Kwang-jib, CEO of Studio Meta-K, provided specific figures for this cost structure. He calculated that a 100-minute film made with premium commercial AI tools could incur a token bill exceeding 500 million won, roughly $366,000. While current prices are kept low by AI companies operating at a loss, Kim projected that a tenfold increase in rates would push costs to 5 billion won, or $3.66 million.

This financial exposure stems from the iterative nature of AI generation. Unlike digital filming, where editing does not cost extra, AI requires payment for every re-render. Kim explained that if a client requests changes individually over ten days, the studio pays for ten separate passes. Batching those changes into a single round lowers the expense, but he argued that without meticulous preproduction, final budgets could exceed those of the celluloid days.

Lee Sang-wook, head of the AI Content Lab at MBC C&I, confirmed that budgeting is difficult because tools update frequently. His lab spent more than a year producing a 139-minute feature titled Raphael, during which new software arrived weekly. Lee noted that a single shot can require ten artists for a full day, and using older technology makes scenes look dated to modern audiences.

Raphael, created by emerging directors with support from the Korean Ministry of Culture, Sports and Tourism, opens in Korean cinemas on November 25. Lee said the theatrical release was chosen to show that AI studios can take commercial risks through IP development rather than just outsourcing work. This strategy follows a 2024 initiative where the lab produced AI shorts that received a cold market reaction.

Cost savings in advertising remain stark. Kim cited a Pizza Hut commercial that originally cost 2 billion won to produce over eight months. His team recreated the spot in two weeks for about 30 million won, cutting expenses by more than 90%. For drama, however, Kim favors a hybrid model that combines live-action star performances with AI-generated background elements like crowds or war scenes.

The two experts differed on open-source tools. Kim praised recent improvements in free models, while Lee remained cautious due to training data transparency and EU AI Act compliance. Lee also noted that delivering AI content to streamers is technically challenging due to strict requirements for bitrate and color grading.

Regarding employment, Kim predicted that large on-set crews will become rarer, with roles like drone operators potentially disappearing. He believes workers who master AI will replace those who do not. Lee compared the shift to the introduction of sound or digital cameras, predicting new roles like previsualization artists and continuity supervisors. He added that actors have generally welcomed AI for handling dangerous scenes, though he maintains that human emotional detail remains irreplaceable.