Application of generative AI in managing projects for 3D game environment development

Abstract

The study focuses on the development of modern technology for creating 3D models of environmental elements for video games using generative neural networks. This approach significantly accelerates the development process of the final product while enhancing the quality and uniqueness of 3D content and ensuring its compliance with the overall design requirements of video games. The article substantiates the feasibility of using artificial intelligence tools for creating prototypes of 3D models of environmental elements and proposes an optimized development technology. The study examines the specifics of the technological process, which includes two main stages: generating models using generative neural networks and refining them further in 3D modeling software. Special attention is paid to the advantages of using generative artificial intelligence for automating the basic stages of development, allowing artists to focus more on detailing, texturing, and animating elements. In particular, the integration of automated tools with traditional 3D modeling approaches is emphasized, improving team efficiency and optimizing resource expenditures. The implementation of new AI tools, such as Sloyd.ai, CSM, and Lumalabs, for creating 3D models that meet the requirements of game engines, is analyzed. The proposed technology was tested in the development of environmental elements for the video game "The Gallery," created as part of an international project. The results demonstrated the effectiveness of combining the speed of AI generation with the quality of manual refinement. The paper outlines the prospects for further development of the technology, which include improving generative neural networks and integrating them with game engines. The proposed approach is promising for optimizing the content creation process and achieving a balance between speed, quality, and product uniqueness.