Capcom Embraces AI: Generating Unique Game Environments

Capcom Leverages Generative AI for Efficient Game Environment Design

Facing escalating video game development costs, game publishers are increasingly turning to AI tools to streamline workflows and reduce expenses. This trend, while controversial, is gaining traction. Activision, for example, reportedly utilized AI for a Call of Duty: Modern Warfare 3 cosmetic item and a loading screen in 2023, sparking debate among fans. EA even declared AI as "the very core" of its operations in September.

In a recent interview with Google Cloud Japan, Capcom's Technical Director Kazuki Abe (known for his work on Monster Hunter: World and Exoprimal) detailed the company's AI experimentation. Abe highlighted the immense time and effort required to generate the "hundreds of thousands" of unique design concepts needed for in-game environments. He cited the creation of assets like televisions, each requiring distinct designs, logos, and shapes, as an example of this massive undertaking. Thousands to tens of thousands of such objects per game necessitate multiple design proposals, complete with illustrations and textual descriptions for communication with art directors and artists.

To improve efficiency, Abe developed a system employing generative AI. This system processes various game design documents, generating design concepts and providing self-feedback for iterative refinement. The prototype utilizes multiple AI models, including Google Gemini Pro, Gemini Flash, and Imagen, and has reportedly received positive internal feedback. The anticipated outcome is a significant cost reduction and potential quality enhancement compared to manual creation.

Currently, Capcom's AI implementation remains focused on this specific system. Other aspects of game development, such as core gameplay mechanics, programming, character design, and overarching narrative ideation, are still primarily handled by human developers.