Adaptive AI Behaviour in Stealth NPCs: A Narrative Review of Current Approaches and Research Directions

Ibrahim Ahmad, Nuraziera Mohd Hatta, Nazreen Abdullasim, Mohd Hafiz Faizal Mohamad Kamil, Norziah Sarudin, and Heru Astikasari Setya Murti
dx.doi.org/10.17576/apjitm-2026-1501-20

Abstract

The problem statement of the current paper is the necessity of creating a dynamic artificial intelligence for the design of stealth non-player characters so that they may create a more dynamic and interesting gaming experience. Today's players expect realistic and flexible game design together with an interesting challenge ratio from the AI techniques applied in the video game, which cannot be done by employing traditional methods based on finite state machines or programming code. In spite of the fact that there has been much progress in reinforcement learning, the development of behavior trees and the combination of the two mentioned techniques which extended the possibilities of AI designs, the synthesis concerning their implementation in stealth NPCs has not yet been done. Thus, the aim of this study is to provide a narrative review of the problem of adaptive AI in stealth NPCs with a focus on the technical and design aspect of development. Literature databases like Scopus, IEEE and ACM digital libraries for the period of 2010-2025 are analyzed.The assessment concludes on the issues and opportunities in the future and it implies that stealth AI research needs to be player-driven assessment systems and cross-genre flexibility. The article introduces a systematic perspective to enlighten researchers and designers in the creation of adaptive NPC behavior in the direction of meaningful and sustainable design of any game.

keyword

Adaptive Artificial Intelligence (AI), Game AI Design, Narrative Review, Player Engagement and Scalability, Stealth Non-Player Characters (NPCs).

Area

Multimedia and Usability