PROMPT ENGINEERING AND LANGUAGE MODELS IN DIGITAL GAMES: A SYSTEMATIC MAPPING STUDY

Abstract

This paper presents a systematic mapping study on the use of Prompt Engineering, Large Language Models (LLMs), and generative artificial intelligence in digital games. The study aims to identify applications, prompting techniques, models, tools, benefits, challenges, and research gaps related to the use of these technologies in game development. The search strategy considered studies published between 2023 and May 2026 and resulted in a preliminary corpus of 35 works, of which 15 were selected for detailed analysis. The results supported the organization of a taxonomy of applications covering interactive narrative and storytelling, dynamic NPCs and dialogues, quests and RPG content, procedural generation of playable content, game design and co-creation, game development pipelines, educational games, and gameplay with LLM-based agents. The analysis indicates that Prompt Engineering acts as a mechanism for specification, control, validation, context retrieval, and integration with game engines. Challenges were also identified regarding narrative coherence, hallucination, latency, computational cost, human authorship, and the lack of standardized evaluation metrics. As a contribution, the paper organizes the analyzed studies and proposes a taxonomy of the identified applications.