AI-BASED NPC INTERACTION SYSTEMS FOR VIDEO GAMES

Abstract

In the era of the fourth industrial revolution, artificial intelligence is rapidly transforming the global gaming industry. Game development has evolved beyond traditional interaction methods, with AI and voice recognition becoming key drivers of innovation in creating immersive player experiences. This graduation thesis explores the development of an AIpowered voice interaction system for NPCs in video games, implemented using Unity Engine and local AI technology (Llama 3.2 3B via LM Studio). The system, built for Windows platforms and powered by C# with Python backend services, features natural voice-driven dialogue, context-aware AI responses, and dynamic NPC behavior routines. The game incorporates advanced features such as speech recognition, intelligent quest management, time-based NPC activities, and a comprehensive economy system. The design aims to demonstrate how AI integration can create more natural and engaging interactions between players and game characters, providing a foundation for next-generation game development approaches. Keywords: Unity, Game 2D, Fantasy Game, AI, NPC.