Red Ram: On-demand Mystery Adventure Game Creation System Using Generative Models

Abstract

With the rise of text/image generation models, there are many efforts to utilize them in digital game production. However, most of these efforts are limited to a small part of the text/image data production process performed by humans. In this research, to verify the possibility of fully automatic creation of game content using generative models, we build a system called Red Ram that creates mystery adventure games on demand. Red Ram automatically creates almost all game content, including the story, character images, and conversation scenarios, based on simple instructions from the user regarding the story. We explain the flow of Red Ram's game content creation using text/image generation models and the system construction procedures. We also discuss the current issues facing Red Ram that we discovered as a result of generating game content.