AI-Powered Communication: A Game Design Approach to Divided Realities

Description

This project explores the potential of AI-native video games to foster rational
discourse in polarized societies. Central to this exploration is the design and implementation of
a game that uses large language models (LLMs) to simulate immersive, real-time conversations
between players and non-player characters (NPCs). The project integrates game design
principles, social theory, and cutting-edge generative AI tools to create a virtual public space
where players engage in dynamic dialogues that influence NPC behavior and narrative
outcomes. Built using Unity and locally hosted LLMs such as Llama 3.2, the system ensures
privacy, responsiveness, and context-aware interactions. AI-generated 2D and 3D assets created
with tools such as Flux Schnell and Trellis enrich the visual and spatial dimensions of the game.
The project critically engages theoretical frameworks from media studies, critical design, and
political communication to explore how digital dialogues can reflect, challenge, and potentially
bridge real-world divisions. Key findings highlight both the promise and current limitations of
generative AI in sustaining coherent, emotionally resonant interactions. While technical
instability and design challenges remain, the project demonstrates that AI-driven games can
serve as platforms for social reflection and engagement. This work contributes to emerging
conversations about the role of artificial intelligence in narrative design, interactive media, and
civic discourse, and provides a foundation for future research and development of AI-powered
public communication tools.
 

Team

Members

Student:

Nick Xinyuan Li 

Supervisor:

Mark Olson