Google used its I/O 2026 event to push Gemini deeper into consumer products, though one relatively quiet announcement carried bigger long-term implications. Project Genie, Google’s experimental AI world model, can now turn real-world locations from Street View into playable virtual environments. The update shifts Genie from a creative sandbox toward an AI-powered world simulation.
Project Genie already allows users to generate explorable environments through text prompts. The latest update adds real-world mapping data into that process. Users can now select real US locations directly within the platform and transform them into stylized, interactive spaces.
Google says users only need to tap the Maps pin option, choose a location, and apply a visual theme such as Ocean World, Stone Age, or Desert Sands. The system then rebuilds that location inside a playable AI-generated environment. Users can also insert custom characters into the world. Google showed examples ranging from animals to comic-style heroes and clay-animation monsters.
The tech giant says the feature relies on ‘Maps Imagery Grounding’ technology. The system uses Street View imagery to help the AI understand physical spaces with higher accuracy.
The announcement may look experimental on the surface, though it signals something larger inside Google’s AI roadmap. Most generative AI products still focus on text, images, or video. Project Genie moves into simulation, which creates a different category entirely.
Google is no longer asking AI to generate static content. It wants AI to build responsive environments that users can move through and interact with in real time. That shifts generative AI closer to gaming engines, virtual production tools, and spatial computing systems.
The rollout remains limited for now. Street View-powered generation currently works only with US locations. Access is also restricted to eligible Google AI Ultra subscribers paying $200 per month.
OpenAI, Meta, and Nvidia are all pushing deeper into AI-generated environments and simulation models. Google doesn’t want to fall behind in that category since it spent years leading the mapping and image recognition technologies.
Project Genie also highlights how valuable Google’s data ecosystem has become in the AI era. Street View imagery, Maps infrastructure, and Gemini models now operate as one combined stack. This combination could eventually matter far more than standalone chatbots. The larger play here is not entertainment alone. Google is building systems that can recreate, simulate, and eventually predict how digital versions of the real world behave.
Also Read: OpenAI Accused of Sharing ChatGPT User Data With Google and Meta