UC Santa Cruz

When we tell stories, we play with them. Theme, character, setting, and plot are some of the main sites of this style of play. Fresh perspectives on these elements allow them to be combined and interpreted and new ways, inspiring further play. Interactive narrative in computation has explored this idea from a number of perspectives, often running into the problem of how to invite and incorporate a naive audience member into a story. In this dissertation, I discuss my work re-examining the assumptions of interactive storytelling, drawing from theatrical practices to create AI systems of expressive performance, their use and impact on narrative design, and ways of supporting the development of experiences built around the idea of players as performers. This combination of design theory, practice, and evaluation blend together into \textit{computational theater}: a new approach to creating interactive narratives and AI driven character behavior rooted in the aesthetics of theatrical performance. Through this approach, I illustrate how performance expands the ways a player can interact with and influence narratives through play with AI controlled characters as well as the resulting implications for the development of AI-based interfaces and simulations.