Fuji television to integrate generative AI across all live-action dramas, potentially cutting vfx work by 70-80%

Fuji Television plans to incorporate generative AI into the production of all its live-action television series starting in October 2027, positioning itself as the first major commercial broadcaster in Japan to fully adopt AI in programming. The company intends to utilize Google AI image generation and editing tools, specifically Nano Banana, for its 16 annual dramas. This integration will be managed by Fuji Television's internal lab, AI Sandbox, which was established in September and is staffed by CG designers trained by the Tokyo production company ANDENT. AI has already been tested, such as generating a bloodied surgical incision scene for the medical drama *24 Hour Emergency Ward*. The new workflow involves training AI on pre-shot footage in a local environment, followed by generating required visuals via prompts, with post-production adjustments made for color and detail. This process can reduce traditional visual effects work by 70% to 80%. Furthermore, the company plans to use AI flexibly for supplementary visual effects and location shooting. Shin'ya Yamazaki of Fuji Television's Technical Art Design Bureau noted that they chose in-house development to avoid increased costs from external vendors. Fuji Television's production strategy manager emphasized that the ultimate decision on AI usage rests on the final quality, stating that the primary goal is ensuring the world depicted in the drama feels authentic and believable.