narrative design

This work addresses the fundamental tension in interactive storytelling between authorial intent and player agency. It proposes Dramamancer, a novel design paradigm for large language model (LLM)-driven interactive narratives that reconciles these competing demands by parsing author-defined story schemata and integrating dynamic response mechanisms to player actions. Leveraging the generative capabilities of LLMs, the system produces narrative content in real time that adheres to underlying story logic while supporting high degrees of interactivity. Empirical evaluation demonstrates that this approach effectively balances narrative coherence with creative control and player autonomy. The study thus offers a reproducible design framework and evaluation pathway for LLM-powered interactive storytelling systems.

Generative AI storytelling faces a fundamental tension between player agency and narrative coherence, compounded by the absence of narrative standards tailored to its generative nature. This paper introduces the first open narrative standard specifically designed for generative AI. Its core innovation is an objective narrative model grounded in authorial intent encoding: creator intentions are formally represented as computable constraints, thereby unifying narrative portability and generation controllability. The methodology comprises four components: (1) formal narrative modeling, (2) semantic encoding of authorial intent, (3) design of generative constraint interfaces, and (4) a cross-platform interoperability protocol. Experimental evaluation demonstrates that the standard preserves 92% narrative consistency while increasing creator controllability over generation by 3.8×. It further enables cross-platform reuse of narrative assets and intent-driven content generation.

This work addresses the challenge of generating executable, state-consistent interactive fiction games using large language models (LLMs). Methodologically, it introduces a story-driven end-to-end framework: (1) an LLM first generates a narrative backbone; (2) a dynamic world state representation is constructed from this story; and (3) executable Python action code is automatically synthesized based on formalized action preconditions and effects, with runtime extensibility to accommodate player improvisation. Its key contribution is the first realization of a closed-loop, tightly coupled “story–state–action” generation mechanism, ensuring both narrative openness and interactive executability. Experiments demonstrate a significant improvement in action code generation success rates and enable players to complete generated games end-to-end, thereby validating the feasibility of LLM-driven, fully automated interactive fiction generation.

This work addresses the long-standing challenge in procedural content generation (PCG) of automatically transforming narrative text into playable visual environments. We propose a lightweight generative pipeline that parses natural-language stories into spatial predicate triplets (Object–Relation–Object), uniquely integrating large language models’ (LLMs) narrative comprehension with GameTileNet’s affordance-aware semantic embeddings to drive semantically grounded 2D tile-based scene layout. Terrain generation across multiple frames is achieved via cellular automata, while object placement adheres to spatial constraints encoded as logical rules. Evaluated on ten diverse narratives, our approach significantly improves semantic object–environment alignment, terrain layering coherence, and cross-frame spatial consistency. Results demonstrate the feasibility and scalability of generating playable, narrative-driven scenes grounded in semantically meaningful, action-oriented representations.

Interactive fiction generation faces challenges stemming from large language models’ lack of world consistency and the limited creative flexibility of symbolic systems. This work proposes IVIE, a neuro-symbolic system that achieves an effective balance between creative freedom and logical coherence for the first time in this domain. IVIE employs a four-stage incremental generation pipeline, delegating creative content generation to a large language model while leveraging symbolic reasoning to verify and constrain the evolving world state. The architecture enables the construction from scratch of fully playable narrative worlds comprising locations, items, characters, and puzzles. Human evaluations demonstrate that the generated content maintains high thematic coherence while significantly enhancing player immersion, thereby validating the synergistic advantage of the approach in reconciling narrative consistency with generative flexibility.

Current LLM-driven game worlds struggle to maintain persistent state consistency, hindering the deep integration of autonomous narrative and simulation. This work proposes a playable virtual world framework governed by a single coordinating agent, modeling game state as a normalized JSON entity tree and formalizing it for the first time as a parameterized-action partially observable Markov decision process (POMDP). Structured action execution and state updates are achieved through a Plan-Diff-Validate-Apply pipeline, JSON Schema validation, and a narrative projection observation mechanism. Empirical evaluation across 15 real-world interaction scenarios demonstrates that the framework effectively supports unrestricted player agency while preserving global world consistency.

This work addresses the challenge of insufficient narrative coherence in purely large language model (LLM)-driven interactive storytelling by introducing a neuro-symbolic architecture that bridges neural and symbolic approaches through a world state transition mechanism. Specifically, Llama 3 70B and Gemini 1.5 Flash are employed to predict state changes within a rule-based system, thereby triggering predefined narrative transitions. This design ensures global consistency of the story world while enabling users to express creative intent through unconstrained natural language input. The system supports both English and Spanish, and user studies demonstrate that the proposed method significantly enhances narrative coherence and fosters more engaging, creative player interaction compared to baseline LLM-only approaches.

This work addresses the challenges of poor coherence, loose structure, and narrative drift commonly observed when large language models generate complex, multi-layered role-playing game content. To mitigate these issues, the authors propose a dependency-driven, multi-stage prompting pipeline that explicitly models narrative dependencies through a structured JSON intermediate representation. The pipeline sequentially generates world settings, non-player characters, player characters, campaign outlines, and quest details, with each stage conditionally informed by prior outputs. By decoupling high-level planning from local elaboration and integrating structured prompts, schema constraints, and explicit data-flow control, the approach substantially enhances logical consistency, structural completeness, narrative coherence, and scalability. Notably, it maintains high-quality output even as content complexity increases.

This study addresses a critical gap in collaborative storytelling research, which has predominantly focused on adult–AI interactions in digital environments, overlooking child-centered approaches that integrate physical play. To bridge this gap, the work proposes the first multi-agent large language model (LLM) framework tailored for children’s physical tabletop gameplay, introducing an iterative Writer–Editor mechanism that enables children to co-create stories through tangible interactions with AI. Experimental results demonstrate that even with minimal rounds of editorial feedback, the system significantly enhances narrative quality, consistently generating age-appropriate, coherent, and creative stories for young users. By extending multi-agent LLM–based co-creation into the domain of children’s embodied play, this research establishes a novel paradigm for child–AI collaborative storytelling.

This work addresses the limitation of existing narrative-driven interactive experience generation methods, which often fail to model story-world consistency, leading to a disconnect between generated content and narrative context. To bridge this gap, we propose explicitly reconstructing a persistent narrative world as the central computational task, formalizing it as a structured representation encompassing entities, locations, semantic relationships, and state evolution. This unified world model jointly supports narrative understanding and interactive content generation. Integrating structured world modeling, narrative semantic parsing, and contextual reasoning, we present the first prototype system that treats persistent world reconstruction as its core objective, enabling automatic transformation from narrative text into playable tile-based environments. Evaluation through three case studies demonstrates that our approach generates semantically coherent and gameplay-consistent interactive worlds, showing promising potential for applications in AI-assisted game design and educational simulations.