Microsoft, in collaboration with Teachable AI Experiences and Xbox Game Studios developer Ninja Theory, has introduced Muse, an advanced AI tool aimed at accelerating game development.

 

 

Microsoft, in collaboration with Teachable AI Experiences and Xbox Game Studios developer Ninja Theory, has introduced Muse, an advanced AI tool aimed at accelerating game development. Muse is designed to generate in-game visuals and player actions by analyzing the 3D game environment.

 

Trained on 1.6 billion parameters, 1 billion images, and 1 billion controller actions equivalent to seven years of gameplay the AI model was developed using user-consented data from Ninja Theory’s game Bleeding Edge. Muse processes visual and controller inputs at a resolution of 300 x 180 pixels, enabling it to generate sequences of up to two minutes and predict future game events.

 

Microsoft's demonstrations highlight Muse’s ability to manage camera transitions, path selection, and character creation while maintaining game physics. The AI is already being used to develop real-time playable models in other games and could also be instrumental in reviving classic video games.

 

To enhance accessibility, Microsoft has released Muse’s model weights, sample datasets, and the WHAM (World and Human Action Model) through Azure AI Foundry. Additionally, Copilot Labs will provide interactive AI-driven gameplay demos later this year.