Google launches Project Genie, AI world builder

Google has launched an experimental prototype called Project Genie, which allows users to create and explore interactive digital worlds generated by artificial intelligence.

The tool is powered by a recently developed world model known as Genie 3. This model is designed to simulate environments, generating the path ahead in real time as a user moves through a virtual space. Users can create a world by typing a text prompt or uploading an image, define a character, and choose how to navigate, such as by walking or flying.

The company stated the project is part of its research into advanced artificial general intelligence (AGI), arguing that such systems must learn to navigate diverse environments.

The initial release is limited. Access is currently rolling out only to users aged 18+ in the United States who subscribe to Google's top-tier AI service, called Google AI Ultra. The company says it plans to expand to more territories later.

Google described Project Genie as an early research prototype with several known limitations. According to the company, the generated worlds may not look fully realistic, might not follow prompts or real-world physics accurately, and can sometimes be difficult to control. Each created world session is currently limited to 60 seconds.