AI & Models
Google integrates Street View into Genie world model
Google DeepMind is integrating Street View data into its Genie world model to enable real-world environment simulation, though the feature remains an experiment with accuracy limitations.
Starting today, Google DeepMind is connecting Street View to Project Genie, the company’s general-purpose world model—an AI system designed to simulate or understand the physical world. The new feature, which enables the simulation of real-world environments, was launched during the Google I/O 2026 developer conference. The integration allows users to generate and manipulate interactive environments anchored to actual geographic locations.
The integration leverages 20 years of Street View data collection, drawing from north of 280 billion images collected across 110 countries and seven continents. Jack Parker-Holder, a research scientist on DeepMind’s open-endedness team, noted that Street View provides imagery from a large portion of the world to help train autonomous agents. Google previously released Genie 3 for research preview in August 2025 and opened access to AI Ultra subscribers in the U.S. in January 2026. Currently, Genie 3 is being used to power Waymo simulators to train self-driving cars on rare events, helping the autonomous driving company scale its simulator beyond the 11 U.S. cities where it has already deployed.
Despite the scale of the data, DeepMind product manager Diego Rivas cautioned that the integration of Street View into Genie is currently an experiment with room for improvement regarding accuracy. The models are not yet physics-aware, meaning they do not understand physical cause and effect. Parker-Holder, a research scientist on DeepMind’s open-endedness team, stated, “I think for this kind of model, it’s maybe six to 12 months behind video in terms of the accuracy and quality, so I think it’s something we will solve.” However, Jonathan Herbert, director of Google Maps, highlighted the model’s spatial continuity—the ability of an AI to maintain consistent environment geometry when changing perspectives. Herbert, who started on the Street View team as an intern 12 years ago, noted that the company has long considered how to build a rich model of the world using Street View data. If a user turns 360 degrees, the AI correctly remembers and simulates the environment behind them.
Why it matters
By anchoring AI world models in massive, real-world datasets like Street View, Google aims to move beyond abstract generation toward functional, interactive simulations for robotics and training.