This study explores how generative AI can improve efficiency in the traditional animation production pipeline. The main goal is to measure how much production time can be reduced by using generative AI to support repetitive tasks such as character design, storyboard-based image production, and animation.
In addition to production speed, the study examines whether preplanned elements—including character design, composition, motion, and camera direction—can remain consistent in AI-generated results. By connecting traditional pre-production materials such as character sheets, storyboards, and previsualization with generative AI, the study explores a workflow that can better preserve the director’s creative intent.
The study also connects AI-based cutscene production with the engineering team’s Camera Control research to examine whether camera movements that are difficult to achieve during generation can be adjusted in post-production. Through this approach, the study evaluates the efficiency, practical potential, and current limitations of integrating generative AI into the animation production pipeline, from pre-production to video generation and post-production.

team
Sanghyuk Moon(speaker) received the B.A. degree in Film and Digital Media
from Konkuk University, Seoul, South Korea. He is currently pursuing the M.A.degree in Immersive Media/Animation Contents at the Graduate School ofAdvanced Imaging Science, Multimedia, and Film, Chung-Ang University, Seoul,South Korea. His research interests include animation production, generative artificial intelligence, and AI-assisted animation content creation.
Yusung Chun received the B.A. degree in Digital Media Design from the College of Arts, Kyungil University, South Korea. [He/She] is currently pursuing the M.A. degree in Immersive Media/Animation Contents at the Graduate School of Advanced Imaging Science, Multimedia, and Film, Chung-Ang University, Seoul, South Korea. [His/Her] research interests include 3D animation, video production,
and AI-based video content creation.
Dahyeon Kye is a Ph.D. student at CMLab, Chung-Ang University, under the supervision of Professor Jihyung Oh. His research focuses on generative models for visual content, including image and video generation, editing, and restoration. He is also interested in 3D/4D generation, world simulation, and interactive generative models that enable users to interact with and construct dynamic visual worlds.