Abstract
Serious games are videogames whose purpose goes beyond mere entertainment. However, serious games use in mainstream education is still limited. The development of serious games is an expensive and complex process that requires the participation of different experts (e.g. domain, educators, graphic artists, and programmers). We consider that new generative AI techniques can help in the prototyping of serious games by reducing and automating some of the processes involved. There is increasing evidence that generative AI techniques such as ChatGPT or GitHub Copilot can increase the productivity of writing or coding tasks respectively. In our case, both prototyping and teaching serious games are complex because of the number and diversity of tasks involved and we are currently investigating whether AI techniques can be used to improve and simplify the process. For example, ChatGPT could support the process of creating the game narrative, and other systems such as Stable Diffusion could ease the creation of some of the graphics resources (e.g., for creating more cohesive and coherent backgrounds). Automating some of the costlier processes of game prototyping can contribute to creating better products by allowing the playtesting of different options for more effective games. As the field of generative AI is in continuous change, this paper presents a working methodology to simplify the development of serious games, that has been instantiated with concrete tools. This working methodology has been piloted effectively by one student from a Master of Design for the development of a serious game and will be tested in a serious games development course. In the article we also explore how these AI techniques can be combined with a game authoring environment such as uAdventure to systematize the development of serious games. The use of Generative AI offers great potential for improving the development of serious games and needs to be further researched alongside its applications for game-based learning education.