Google Launches Project Genie AI Virtual Worlds Tool

Google has launched Project Genie, an experimental tool that enables users to create and explore interactive virtual environments generated from simple text prompts. The prototype, powered by the company's advanced AI models, became available on Wednesday to subscribers of its premium Google AI Ultra plan in the United States, marking a significant step in making sophisticated world simulation technology accessible beyond internal testing.

Project Genie builds on Genie 3, a general-purpose world model that Google DeepMind unveiled in August. This AI system simulates dynamic environments by predicting how scenes evolve and respond to user actions, generating content in real time without relying on prebuilt assets or hardcoded physics engines. Users interact with the worlds through a web-based prototype, navigating via keyboard controls such as WASD keys to move a customizable avatar. The tool renders environments at 1280 by 720 pixel resolution and up to 24 frames per second, allowing sessions of up to 60 seconds.

To begin, users enter natural language descriptions into two input fields: one for the overall 3D environment, such as a volcanic wasteland or an enchanted forest, and...