en.Wedoany.com Reported - Google announced at the I/O 2026 developer conference that its 3D world generator, Project Genie, has added a new feature: it can now access Google Street View imagery to generate explorable 3D environments of real locations, making AI-generated worlds closer to real-world scenes.

Users can select a U.S. location via a map icon within Project Genie, then choose a style setting such as "Desert World" or "Stone Age," and describe a game character. Genie will generate a fantasy world based on the Street View images. Google's official blog used the Golden Gate Bridge as an example, showcasing an underwater world where users can play as a diver to explore. The technical foundation for this feature is the image processing tool "Maps Imagery Grounding," launched in April this year, which anchors generative AI images into the real road perspectives of Google Street View. Google stated that this feature currently only supports locations in the United States, with other regions to follow later.

Project Genie was previously only available to users in the United States. With this announcement, Google is gradually opening Project Genie to regions outside the U.S., but users still need to subscribe to the premium plan "Google AI Ultra." At the developer conference, the original $250 per month plan was replaced by two new Ultra plans ($100 and $200 per month). Project Genie remains in the more expensive $200 plan, though the price has decreased.

Project Genie is based on Genie 3, Google's general-purpose world model capable of generating diverse and limited-interaction environments. Google's purposes for researching such world models include virtual training for autonomous systems such as AI agents, robotics, and self-driving cars. In January this year, Google first opened the model to users and named this research direction Project Genie. Genie is developed by Google DeepMind, while Project Genie is managed by Google Labs.

Google refers to Project Genie as an experimental prototype, and its functionality remains significantly limited: the generated 3D environments can currently only be explored for about a minute, and interactivity is extremely limited. To generate a world, users only need text input, images, and now, real locations can also be used. Following the January announcement, shares of major game companies came under pressure, as investors viewed AI models like Project Genie as potential competitors to traditional game development, but actual testing shows the industry has nothing to worry about for now.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com