Google Unveils Project Genie, an AI Tool That Lets Users Build and Explore Their Own Virtual Worlds

Google has introduced Project Genie, an experimental artificial intelligence web application that allows users to design and explore their own interactive virtual environments. The new tool, announced by CEO Sundar Pichai, marks another step forward in Google’s efforts to push the boundaries of generative AI beyond text and images into fully explorable digital worlds.

Sharing his early experience with the platform, Pichai revealed that he had already spent time testing the tool himself. Describing it enthusiastically, he wrote, “Project Genie is a prototype web app powered by Genie 3, Nano Banana Pro + Gemini that lets you create your own interactive worlds.” He added, “I’ve been playing around with it a bit and it’s out of this world:)”.

At the heart of Project Genie is Genie 3, Google’s world-generation AI model first introduced last year as a limited research preview. The system works alongside Gemini and Nano Banana Pro to create visually rich and responsive environments. Together, these models enable users to craft scenes, characters, and landscapes simply by typing prompts or uploading images. Even everyday objects, such as toys or sketches, can be transformed into elements of a virtual setting.

The process begins with what Google calls “World Sketching.” Users describe what they want to see, and the AI instantly generates an environment that can be explored from first-person or third-person perspectives. Unlike static AI images, these spaces are interactive. As users move around, the system renders new parts of the world in real time. Google says the model also simulates physics and object interactions to make the experience feel more lifelike.

Project Genie also includes a remix feature. Instead of building from scratch, users can start with pre-made worlds available on the platform and customize them with their own ideas. This lowers the barrier for beginners while still allowing advanced creators to experiment freely.

However, the current version has some limitations. Each generated world can only be explored for 60 seconds at a time. Google also cautions that visuals may not always appear completely realistic and users might occasionally experience delays in character controls.

For now, access to Project Genie is restricted to Google’s AI Ultra subscribers in the United States. According to Google DeepMind’s Diego Rivas, the release is meant to help the company understand how people might use the technology in real-world scenarios. He explained, “It’s really for us to actually learn about new use cases that we hadn’t thought about.”

Looking ahead, Google believes the technology could have applications beyond entertainment and creativity. The company suggests that similar AI-generated simulations might one day help robots learn to navigate physical environments more effectively.

With Project Genie, Google is signaling that the future of AI may not just be something users read or watch—but something they can step inside and explore.