Google DeepMind Integrates Street View into Project Genie AI

Google DeepMind integrated Google Street View imagery into Project Genie, an AI-powered world model capable of generating interactive 3D environments from text prompts. Announced at the Google I/O developer conference, the integration utilizes a dataset of 280 billion images across 110 countries to anchor simulations in real-world locations, though the feature is initially limited to the United States.

Users can transform real-world sites using imaginative styles, such as depicting the Golden Gate Bridge in an underwater setting. The technology is currently rolling out globally to adult Google AI Ultra subscribers. While the model provides strong spatial continuity, officials noted it remains an experiment and lacks full physics awareness.

Beyond consumer creativity, the model serves educational purposes and robotics development. Waymo uses the technology to power simulators that prepare self-driving cars for rare events. Research scientists emphasized that combining real-world imagery with world-simulation capabilities creates a powerful tool for both robotic agents and human interaction.