Microsoft unveils AI model to revolutionise video game testing

Microsoft experts have developed an artificial intelligence (AI) model designed for testing and optimizing video games.

Named World and Human Action Model (WHAM), the AI model is capable of both generating graphics for video games and simulating player actions, running complete gaming sessions. This innovation aims to save valuable time during video game testing, Caliber.Az reports via foreign magazine. 

WHAM was trained on 1.6 billion parameters, utilizing 1 billion images and gaming controller actions, which equals approximately seven years of continuous gameplay. The model was specifically trained using Bleeding Edge, a game developed by Ninja Theory. WHAM is able to generate gameplay, predict game progression, and understand player behavior. It also comprehends 3D gaming environments and game physics, enhancing its ability to simulate realistic gaming experience.

The World and Human Action Model (WHAM) is an innovative artificial intelligence (AI) model developed by Microsoft to streamline video game testing and optimization. WHAM was created to simulate both the actions of players and the interactive environments within video games. It is designed to generate in-game graphics, predict gameplay progression, and analyze player behavior in real-time, significantly reducing the time and resources traditionally required for game testing.

By Naila Huseynova