Chris Bremble, the founder of Base Media, a visual effects and animation firm based in Beijing, and an Emmy award winner, recalls his first encounter with generative artificial intelligence. He discovered it at an industry conference in 2018, where he witnessed a software demonstration that could replicate smoke from a photograph in just five seconds. This experience caused Bremble to realize the inevitable impact of such technology and feel a sense of urgency. While generative AI has yet to be widely adopted by the movie post-production sector, it has already sparked concerns among actors and writers in Hollywood about its influence on their craft and job security, contributing to the current strike. China, where Bremble maintains offices in several cities, has already witnessed layoffs in the video gaming industry due to generative AI. To stay ahead in this technological landscape, Base Media has collaborated on major Hollywood productions like "Black Panther: Wakanda Forever, " the "Star Wars" limited series "Obi-Wan Kenobi, " and Chinese films such as "The Wandering Earth. " Base Media recently partnered with Chinese streaming giant iQiyi to explore the potential of generative AI. Specifically, Wonderworks, iQiyi's children and animation division, wanted to test the AI's capability to generate 3D locations, as this is the most labor-intensive and costly aspect of animation production. Building entire virtual settings requires significant efforts, including constructing storefronts, windows, castles, and city corners. Normally, 30% of the animation production is dedicated to creating locations. Wang Huiyu, head of Wonderworks, provided Base Media with a 90-second script about a girl searching for her friends and the iQiyi animation series called "The Roofus, " which spanned 260 minutes. They fed the entire series into the AI system and tasked it with replicating the animation style and appearance desired by Wang. Igor Lodeiro, Base Media's visual effects supervisor in their Xiamen office and a daytime Emmy winner, led the project. Initially, Lodeiro had low expectations for the AI tools but was surprised by the results. The team started with a rough cut, manually animating the lead character and placing gray blocks as placeholders. They then used photos and keywords to guide the AI in generating the appropriate visuals.

Although there were some initial anomalies and distortions, the frames of the test video closely resembled "The Roofus" show. In the final video, the animated character skateboarded across a wooden bridge in Beijing, traversed curved Chinese rooftops, and hung from a gargoyle overlooking the sprawling city of Paris. AI generated every location and even created the skateboarding motion. The video required three weeks to complete, compared to the over eight weeks it would have taken using traditional human animation methods. The integration of AI is seamless, making it challenging for the audience to discern the AI-generated aspects. Bremble appreciates this, emphasizing the importance of preserving the illusion that the audience is not consciously aware of AI's involvement. For him, AI's potential lies in enabling top talent to accomplish more in a day and providing greater predictability in production scheduling. Lodeiro envisions AI taking over more tedious tasks such as storyboarding, leading to better quality animations and the creation of unique visuals that would be difficult to achieve with conventional methods. He is genuinely excited about AI and believes it enhances rather than threatens job security. Base Media, however, understands that not all clients are receptive to AI integration. While their Chinese clients eagerly embrace AI's potential for increased productivity, their U. S. clients demand assurances that no AI has been involved in the creation of their work. Concerns about copyright infringement arise among Chinese content creators, but they are also excited about AI's ability to enhance creativity in a fiercely competitive environment. Wang, for instance, imagines AI's impact on making animated food appear more realistic, with intricate textures, steam, and transparency. She hasn't explored AI's potential with natural elements like flower petals and dew but plans to conduct further trials with the goal of eventually producing an entire movie using generative AI.