Narrative Adherence in LLM-driven Games
Term
4. term
Education
Publication year
2025
Submitted on
2025-06-03
Pages
14
Abstract
This paper presents ChatRPG v3, a narrative-driven AI game master for single-player role-playing games. Building on prior versions, we introduce SENNA, a five-agent system that enables adherence to structured, pre-written adventure modules while preserving player agency. The system uses a narrative graph to track story progression. Furthermore, we design redirection strategies to guide players back to the intended narrative arc when deviations occur. We evaluate the effectiveness of SENNA and these strategies through a user study combining live gameplay, alternative scene comparisons, and player feedback. Results show that structured redirection techniques rooted in world logic, such as NPC influence and in-world consequences, enhance narrative coherence without diminishing immersion or autonomy. Our findings expand the design space for AI-driven interactive fiction and offer practical insights for balancing authored stories with emergent player input.
Keywords
Agentic AI ; AI ; AI Game Master ; Artificial Intelligence ; C# ; ChatRPG ; ChatGPT ; Choose Your Own Adventure ; CYAO ; DM ; GM ; Dungeons and Dragon ; Dungeon Master ; Game Master ; Games ; Generative AI ; Narrative Adherence ; Narrative Redirection ; GPT-4o ; Interactive Fiction ; LangChain ; Large Language Models ; LLM ; Multi-Agent System ; Narrative ; OpenAI ; Player Experience Inventory ; PXI ; Artificial Social Agent Questionnaire ; ASAQ ; Prompt ; Prompting ; Prompt Engineering ; Role-Playing Games ; RPG ; Storytelling ; TTRPG ; User Engagement ; Narrative Graph ; ReAct ; SENNA ; Redirection Strategies ; Redirection ; Pre-written Adventure
Documents