A Multi-dimensional Framework for Artificial Intelligence Integration in Game Development across Artistic, Educational, and Environmental Contexts​

Abstract

This research investigates artificial intelligence integration patterns in game development through systematic analysis of diverse implementations across multiple application domains. Within this investigation, artificial intelligence encompasses both traditional computational approaches—including rule-based systems, algorithmic methods, and procedural generation techniques—and contemporary machine learning technologies, reflecting the broad spectrum of computational methods that enable intelligent behaviors in interactive systems. Employing a grounded theory methodology, this study examines six case studies that span three primary technological approaches — Procedural Content Generation, Intelligent Agent Systems, and AI-assisted Interaction Frameworks — applied within artistic expression, educational training, and environmental awareness contexts. The empirical investigation includes: (1) sensor-based NFT generation systems; (2) parameter-driven PCG for environmental art installations with real-time climate data integration; (3) physiological signal processing integration with AI-driven agent generation systems; (4) reinforcement learning agents modeling ecological dynamics in cultural contexts; (5) adaptive computer vision for educational applications with player-in-the-loop error correction; and (6) large language model integration for philosophical education. Through systematic analysis of these diverse implementations, consistent patterns emerge that reveal two primary organizational dimensions for understanding AI integration in game development. The primary contribution of this research is the identification and documentation of these emergent patterns, which inform a multi-dimensional classification framework that moves beyond isolated technical implementations to demonstrate how AI technologies can be strategically combined and contextualized to address specific design objectives. This empirically-grounded framework provides developers with evidence-based guidance for technology selection and integration strategies, advancing our understanding of how computational intelligence can enhance both player experiences and societal outcomes through interactive digital media.

Date of Award	2025
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Kang ZHANG (Supervisor) & David Kei Man YIP (Supervisor)