Why Are We Like This?
Melanie Dickinson, Max Kreminski, Michael Mateas, and Noah Wardrip-Fruin
Why Are We Like This? (WAWLT) is an AI-augmented digital story construction and collaborative, improvisational writing game in which two players write a story in a pastiche of the cozy mystery genre, with support from a simulation-based AI system that operationalizes character subjectivity.
WAWLT explores how computation can enable new forms of playful, social creative writing practice. By running, querying, and updating an underlying storyworld simulation, the AI system provides players with inspiration and keeps the story moving forward, even when the players are unsure what should happen next. Players collaboratively select author goals they would like to work towards throughout the story, and select actions for characters to perform, either from a set suggested by the system or by querying the action possibility space in a custom story sifting interface. The suggested actions are continually reassessed (using simulation rules) based on what each character might want to do next, prioritizing actions that could fulfill the current author goals. Whenever players select an action to be performed in the storyworld, its effects are realized in the simulation, and a generated action description is appended to a textual transcript recounting the story so far, which players freely edit as the story develops.
The system uses the newly developed technology of story sifting - the extraction of narratively potent sequences of events from the chronicle of all the events that have taken place within a simulation. Sifting is used by players to guide the story, and used to implement character subjectivity: