Autonomous Narratives: Fusing Traditional Narratological Theories with AI Methodologies for Controlled, Character-Driven, Automated Story Generation and Analysis.

GTA funded PhD studentship in Computing

Highlights

1. Controlled Character Narrative: The Intersection of Generative Agents and Narratological Theories.
2. Trade-off Analysis: Understanding the Balance between Control and Agency.
3. Story Analysis and Comprehension: Leveraging Generative Agents Potential in Narrative Examination.

Project

This research aims to bridge the gap between traditional narratological theories and advanced artificial intelligence (AI) methodologies, with a central focus on crafting controlled, character-centric automated stories. The project proposes the innovative utilization of AI-powered generative agents, based on the use of large language models, as characters within the narratives, with each agent initialized to reflect the depth and nuance of a proper story character.

Integral to this study is a nuanced balance between the control provided by established narratological structures (e.g., the Freytag’s Pyramid, the Hero’s Journey, or Propp’s morphology for folktales) and the agency offered to the character-driven generative agents. By grounding agent interactions and narrative progression within pre-defined narratological morphologies, the research ensures that the emerging stories align with accepted narrative principles and are inherently appealing to human readership.

The character agents, capable of autonomous decision-making within the limitations of their pre-determined arcs, create an interesting blend of predictability and surprise, contributing to the richness and dynamism of the narrative. This unique blend presents an opportunity to delve deeper into understanding the trade-off between controlling narratological structures and granting agency to generative agents.

In addition to improving storytelling and story generation procedures, the research aims to exploit the capabilities of AI to dissect and understand existing stories; the use of generative agents with different characteristics allows for a multi-perspective analysis approach.

The research project outlined above can have implications across several fields. Within the entertainment industry, applications could stretch from video gaming, where the study could lead to more complex, character-driven, and immersive gaming experiences, to film and television, potentially creating AI-powered scripts and interactive narrative experiences. For businesses, an AI capable of crafting compelling narratives could be valuable in the sector of advertising and marketing strategies and could enhance the increasingly popular chatbot technologies.

PhD start date 23 September 2024

Enquiries to project supervisor Dr Marco Volpe  mv163@leicester.ac.uk    or cmspgr@le.ac.uk

Further details and application advice at https://le.ac.uk/study/research-degrees/funded-opportunities/cms-gta