End-to-End Game Development with Contemporary AI Systems

This project examines the current (2025/2026) capabilities and limitations of contemporary artificial intelligence (AI) systems for developing a complete videogame from initial concept to a finished, playable product. The study focuses on (a) the practical feasibility of using AI tools across all major stages of game development, (b) the roles AI systems actually play within the game-development pipeline, and (c) the suitability of different classes of AI models for different types of tasks. For this study, “AI systems” include large language models, code-generation tools, image and video diffusion models, music and sound-generation systems, and other generative or assistive tools in common use in 2025/2026. “All aspects of a game” refers to core production activities such as game design and documentation, narrative and dialogue, code and technical implementation, visual assets, audio and music, quality assurance and balancing, and supporting materials (e.g., store pages or marketing copy). The study is primarily exploratory and descriptive: the goal is to document workflows, catalogue which tools are most effective for which tasks, and identify recurrent limitations and failure modes when attempting AI-supported end-to-end game development.