Research Article
AI Use in Game Design: Optimizing NPC Interaction While Preserving Narrative Stability
@INPROCEEDINGS{10.4108/eai.22-5-2026.2365234, author={Hau Tak Zeng}, title={AI Use in Game Design: Optimizing NPC Interaction While Preserving Narrative Stability}, proceedings={Proceedings of the 4th International Conference on Image, Algorithms, and Artificial Intelligence, ICIAAI 2026, 22-24 May 2026, Singapore, Singapore}, publisher={EAI}, proceedings_a={ICIAAI}, year={2026}, month={8}, keywords={Artificial intelligence; Game design; NPCs; Large language models; Narrative stability}, doi={10.4108/eai.22-5-2026.2365234} }- Hau Tak Zeng
Year: 2026
AI Use in Game Design: Optimizing NPC Interaction While Preserving Narrative Stability
ICIAAI
EAI
DOI: 10.4108/eai.22-5-2026.2365234
Abstract
Artificial intelligence is becoming more important in game design, especially in NPC interaction. LLM-driven NPCs can respond to players in real time and make dialogue more open and flexible than traditional scripted systems. However, this openness also creates design risks, including character inconsistency, narrative deviation, and reduced control over story structure. This paper examines how AI-driven NPC systems change narrative design in games and focuses on one central issue: how to preserve narrative stability while expanding interaction. By comparing scripted NPCs, Detroit: Become Human, PANGeA, and evidence from educational and serious games, this paper argues that hybrid design is more practical than unrestricted generation in narrative-heavy games.
