Video game development using artificial intelligence

This bachelor thesis explores the impact of generative artificial intelligence on the transformation of the video game industry and empirically verifies the efficiency of its implementation compared to traditional development methods. The aim of the thesis is to evaluate the benefits and limitations of AI across three key game components: game logic (code), graphics (textures), and sound. Within the practical part, two fully functional versions of a 2D pixel art platformer were created in the Unity game engine. The first variant was developed using a traditional approach, utilizing manual programming in C#, drawing in the GIMP software, and public sound libraries. The second variant maximized the use of generative AI tools, specifically the Google Gemini, PixelLab, and ElevenLabs models. For the final evaluation, the Analytic Hierarchy Process (Saaty's AHP method) was used, synthesizing objective technical measurements (time tracking via timesheets, code performance in the Unity Profiler) with a subjective evaluation of the aesthetic quality of graphics and sound through a questionnaire survey. The results show that AI methods dominate in speed and visual complexity when creating graphics and animations. In the area of programming, AI significantly accelerates the creation of architecture; however, it shifts the developer's role from a creator to an editor who spends most of their time integrating and debugging logical errors. Conversely, in the field of audio production, the traditional approach proved to be more flexible and efficient. In conclusion, the thesis formulates recommendations for smaller game studios and independent developers in the form of a hybrid development model that optimizes the extent of delegating tasks to AI.