Abstract
This study focuses on how to make use of generative AI in game level development by procedurally populating assets at a code-based level in a video game. The researcher aims to enhance developers’ efficiency in aesthetic milestone by adapting a pipeline in which generative AI handles decoration process that has less impact on gameplay. Consequently, the researcher creates a pipeline outlining the best practices for conducting the pipeline using generative AI. The researcher constructed a game level in Left 4 Dead 2 to explore the effectiveness of this pipeline. Testers play the level and provided feedback regarding their experiences, especially on decoration quality. The researcher analyzed this data to confirm or deny whether the pipeline enhances efficiency and at the same time maintains the decoration quality of an average human-designed level.
Degree Date
Spring 2025
Document Type
Thesis
Degree Name
M.I.T.
Department
Level Design
Advisor
Mike Porter
Acknowledgements
A special thank you to Professor Mike Porter for his invaluable guidance and support throughout the development of this research. Gratitude is also extended to Professor Katie Wood Clark and Professor Karl Steiner for their insightful feedback and encouragement during the process. Finally, sincere appreciation goes to all the playtesters whose participation and feedback contributed significantly to the evaluation of this study.
Format
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License
Recommended Citation
Jiang, Tianxing, "Using Generative AI to Procedurally Populate Assets in an ASCII Level" (2025). Level Design Theses and Dissertations. 18.
https://scholar.smu.edu/guildhall_leveldesign_etds/18
