<![CDATA[The rapid evolution of digital media has increased the demand for adaptive storytelling approaches capable of sustaining audience engagement beyond conventional linear narratives. This study investigates the integration of Large Language Models (LLMs) into interactive non-linear storytelling to enhance audience immersion and narrative personalization. A conceptual AI-driven narrative framework is proposed by incorporating adaptive weighting mechanisms inspired by modified gravity theory to dynamically adjust narrative progression according to user interactions and contextual preferences. The proposed framework is evaluated through a comparative analysis between conventional branching narratives and AI-driven adaptive narrative structures using engagement-oriented performance indicators. The findings indicate that adaptive AI-based storytelling improves narrative flexibility, emotional engagement, and user agency compared with traditional fixed-choice approaches. Rather than treating narrative progression as a static sequence, the proposed framework dynamically modifies story development based on real-time interaction, enabling a more personalized storytelling experience. The study contributes to the advancement of intelligent media broadcasting by introducing an interdisciplinary framework that combines artificial intelligence with adaptive narrative modeling. These findings provide practical implications for interactive entertainment, digital education, and AI-assisted content creation while offering a foundation for future empirical research on adaptive storytelling systems.]]>
