Microsoft received 3 granted patents across AI & Machine Learning (1), Cloud Gaming (1), and Networking (1).
The AI & Machine Learning patent covers using generative models to create narrative content from natural language prompts and adapt games based on player engagement. In Cloud Gaming, Microsoft patented a system for displaying personalized ads and branding during low-interaction moments without altering the underlying game code. The Networking patent describes technology for enabling local split-screen multiplayer in online-only games by running multiple instances and routing inputs to simulate couch co-op on a single display.
Microsoft's single AI and machine learning patent describes technology where generative models work as a development assistant, converting natural language descriptions into actual game content while simultaneously adapting that content during gameplay based on how players interact with it. The system creates a two-way loop where designers use AI as a creative partner during production, then the same AI continues generating and evolving content in real time as players engage with the game. Player feedback metrics feed back into the content generation process, turning players into indirect co-creators of their own experience.
The cloud gaming patent addresses the challenge of personalizing and monetizing streamed games without modifying their source code. Microsoft's approach analyzes player interaction patterns to predict when engagement naturally dips, then displays overlay content like advertisements or branding during those moments. By layering personalized elements on top of the video stream rather than injecting them into the game itself, the system works with any cloud-streamed title regardless of whether developers designed it for such features.
In Networking, Microsoft patented a workaround for the disappearing split-screen feature in modern games. The technology runs multiple instances of an online-only game and treats them as separate network clients, then combines their outputs on a single display while routing controller inputs to the appropriate instance. This approach recreates couch co-op experiences without requiring developers to implement native split-screen support or needing powerful enough Hardware to render multiple viewpoints simultaneously.