New Frontiers in AI for Game Design (Dagstuhl Seminar 25292)

Abstract

Game design has often influenced, and been influenced by, computer science research. In recent decades researchers and designers have sought to bring these two fields even closer together: to find new ways to think about the game design process; new ways to drive innovation in computer science through playful exploration; and ultimately find new ways to play, design and think about games through computational lenses. AI is impacting the creative industries in more ways than ever before, some welcome, others less so. It is important to find ways for both researchers and practitioners to come together to map out possible futures for this space, to understand where research can contribute, what it can learn from game design in return, and how we can enrich the creative practice of everyone involved.
This report covers Dagstuhl Seminar 25292: New Frontiers for AI in Game Design. It outlines the motivations for organising the seminar, summarises many of the working groups that took place, and disseminates some of the games, theories and other materials created during the seminar. The report offers theoretical frameworks, working prototypes and exploratory discussions that present many possible futures for both the creative practice of game design, and the academic field of games research. None of these futures are singularly correct, and many more remain out there to be found; this document merely charts out some possible paths into the unknown that we found exciting to consider.