Shape of the Maker : The Representation of Artificial Intelligence in the Horizon Video Game Series

Abstract

This thesis examines how the Horizon video game series by Guerrilla Games constructs its artificial intelligence figures as narrative subjects, and what this construction reveals about contemporary cultural thinking on AI, human accountability, and the capacity to feel. The primary research question asks how the series uses narrative and formal strategies to build AI interiority, and what different relationships to feeling: its presence, absence and suppression, produce in narrative and ethical terms. Two subsidiary questions address what the series argues about human responsibility toward AI, and how it uses mythology, both in the designed naming of its AI system and in the received cosmologies of its fictional tribes, to explore what happens to technology when it is forgotten and made sacred. The research method is close reading, applied to the games' data logs, hologram recordings and dialogue encounters as narrative texts. The theoretical framework draws mainly on AI narrative theory, affect theory, feminist AI scholarship and game narratology. The data consists of four games in the series released between 2017 and 2023. The findings reveal that the series consistently constructs AI interiority as a relational rather than internal phenomenon. The capacity to feel is not a fixed property of AI systems but something produced, permitted or suppressed by the quality of the human relationships surrounding them. Every catastrophe in the series is traceable to a specific human failure rather than simple machine awakening or autonomous goal-seeking. This subverts the dominant AI uprising narrative. The series's mythological registers, both the classical naming architecture of its AI and the cosmologies built around it by the fictional tribes, are read as continuous with this argument. The thesis concludes that the Horizon series offers a coherent and analytically significant alternative to many dominant AI narrative templates. It is one in which the shape of the AI reflects the quality of the human relation that made and governs it, and in which feeling functions not simply as a humanizing aspect but as the condition under which a powerful system can be trusted.