Player-Driven Emergence in LLM-Driven Game Narrative

We explore how interaction with large language models (LLMs) can give rise toemergent behaviors, empowering players to participate in the evolution of gamenarratives. Our testbed is a text-adventure game in which players attempt tosolve a mystery under a fixed narrative premise, but can freely interact withnon-player characters generated by GPT-4, a large language model. We recruit 28gamers to play the game and use GPT-4 to automatically convert the game logsinto a node-graph representing the narrative in the player's gameplay. We findthat through their interactions with the non-deterministic behavior of the LLM,players are able to discover interesting new emergent nodes that were not apart of the original narrative but have potential for being fun and engaging.Players that created the most emergent nodes tended to be those that oftenenjoy games that facilitate discovery, exploration and experimentation.