Introducing StreetReaderAI: a new and more accessible street-level imagery prototype. This innovation uses context-aware, real-time AI and accessible navigation controls to redefine the immersive streetscape experience, making it more inclusive for everyone. Developed with blind and sighted accessibility researchers, StreetReaderAI leverages multimodal AI and Gemini Live to provide real-time AI-generated descriptions, dynamic conversations, and accessible navigation. Our study with blind screen reader users showed positive reactions, highlighting the significant accessibility advancement. Check out the blog post to learn more: https://goo.gle/47sOvJc
About us
From conducting fundamental research to influencing product development, our research teams have the opportunity to impact technology used by billions of people every day. We aspire to make discoveries that impact everyone, and sharing our research and tools to fuel progress in the field is fundamental to our approach.
- Website
-
https://research.google/
External link for Google Research
- Industry
- Technology, Information and Internet
- Company size
- 1,001-5,000 employees
Updates
-
Google Research reposted this
The most annoying part of city driving is the constant stop-and-go traffic 🚘 But what if we could use AI to sync up lights and make commutes significantly smoother – and do it in a famously congested city like Boston? I know Boston’s tricky streets firsthand from my time in grad school, and I’m thrilled that our Project Green Light initiative has not only hit a major milestone in the city but is delivering real impact on the ground. For the first time, our AI-powered recommendations live at over 10% of intersections in a major US city - 114 intersections in Boston and counting. Here's a look at the impact and the genius behind the approach: 🚘 Up to 33% Reduction in Stop-and-Go Traffic: In Boston, the use of Project Green Light has cut down stop-and-go traffic by up to a third! This means drivers are significantly less likely to wait more than one red light cycle to get moving. 🚘 Emissions Cut by Up to 10%: By minimizing unnecessary braking and accelerating, the project has the potential to reduce intersection emissions by up to 10%. 🚘 AI for Faster Implementation: Our AI models analyze Google Maps driving trends to model traffic patterns and generate signal timing recommendations for city engineers. Critically, our city partners tell us they can implement an AI-driven Green Light recommendation in as little as five minutes, dramatically speeding up a process that can take weeks per intersection without the tool. 🚘 The Global Scale: Boston is part of a growing list of over 20 international cities on four continents where Project Green Light is helping to optimize traffic and reduce pollution. This technology is a powerful testament to how smart, scalable AI can directly improve quality of life and deliver a measurable climate benefit—not just in the digital world, but on our city streets. Check out Mayor Michelle Wu’s announcement on the project's expansion in Boston below. Announcement earlier in June: https://lnkd.in/gZkH5vjH Learn more about Project Green Light: https://lnkd.in/gghm5Gwy
-
We’ve launched our new AI-powered @Fitbit personal health coach, which was developed thoughtfully, iteratively and grounded in scientific research. Our development is grounded in decades of science, incorporating feedback from a Consumer Health Advisory Panel and professional coaches. We validate its helpfulness, accuracy, and safety using the SHARP evaluation framework, backed by millions of human annotations. Read more about our approach and additional studies that helped shape the coach → goo.gle/4qwfQD2 This commitment to rigor is paramount. Join the Public Preview and help us shape the future of adaptive wellness → https://lnkd.in/g8JTrTzw
-
-
Our VP & GM of Google Research, Yossi Matias, posted about how Google Research’s approach is helping solve challenges in three areas that affect so many people — check it out ↓
Over the past two weeks, we've seen a flurry of new results from Google Research, from genomics to quantum computing to geospatial understanding. These breakthroughs exemplify what I call the “magic cycle” of research: addressing global challenges and opportunities with foundational research that leads directly to real-world applications and solutions. These solutions not only benefit millions around the world, but they also uncover more important problems to solve. Research is our chance—and imperative—to improve day-to-day lives, and address societal challenges and opportunities, and it means that research and innovation is never 'done.' Three highlights on latest from Google Research in collaboration with other teams: ✨ Fighting Cancer with AI Our decade-long effort in genomics led to DeepSomatic, an AI-powered tool that accurately spots genetic variants in cancer cells. Our partners used it to identify 10 new genetic variants in childhood leukemia samples that were missed by previous techniques, advancing personalized treatment. Separately, the Cell2Sentence-Scale 27B model successfully generated a novel, validated hypothesis for a combination cancer therapy. ✨ Verifiable Quantum Advantage We demonstrated the Quantum Echoes algorithm on our Willow chip. This marks the world's first verifiable quantum algorithm that brings us a step closer to practical applications of quantum computing, such as designing better materials and medicines, by more precisely modeling quantum behavior that classical supercomputers cannot. ✨ Understanding Earth Our new initiative, Earth AI, focuses on synthesizing disparate geospatial information and predictive models. This multi-year effort will make it possible to tackle complex planetary science and crisis resilience questions—like precisely predicting vulnerable communities and infrastructure during an impending storm, or help in complex logistics planning—that are currently impossible to answer due to the complexity of integrating diverse data. The magic cycle is accelerating significantly, powered by stronger models, and agentic tools, like AI co-scientist and AI-based expert-level empirical software. And it applies across many disciplines and domains. Today, the energy was palpable at our flagship Research@ event in Mountain View. The opportunity to advance research and AI is greater than ever and I’m excited about how AI can be an amplifier of human ingenuity. Energized by the amazing team of Google Research and what they can do in collaboration with great colleagues and the community. Read more in my blog: https://lnkd.in/d778XcWm
-
-
Missed us on Tuesday? Here's another chance to experience Android XR firsthand at the #ICCV2025 Google booth today at 11:30am HST! See the advanced Perception stack for seamless immersive experiences, translating vision research into real-world XR products. https://lnkd.in/gyjCEkRT
-
As our Chief Scientist of Quantum Hardware, Michel Devoret, notes, the field of quantum mechanics is thrilling because "nature has been clever, more clever than you." In this video about Google Quantum AI’s latest breakthrough algorithm, Quantum Echoes, he discusses how his pioneering work on superconducting artificial atoms is the basis for the Willow processor chip. That powerful foundation enabled this verifiable result—a crucial step toward real-world applications in fields such as chemistry and materials science. Learn more about our latest breakthrough → https://goo.gle/42PG6y8
-
Today, we are sharing our latest advancements for Google Earth AI — a fundamental leap in planetary understanding achieved through specialized AI models and agentic reasoning. This includes new research on Geospatial Reasoning and expanding access to our Earth AI models on Google Earth and Google Cloud. Geospatial Reasoning is a framework that, using Gemini models, lets AI automatically connect different Earth AI models — such as weather forecasts, population density, and satellite imagery — to answer complex, multi-modal questions. 🌍 Connecting the Dots: Geospatial Reasoning is now available for Trusted Testers. For instance, the nonprofit GiveDirectly is using it to combine flood and population density information to identify those who need direct aid most during a disaster. 🌍 Deeper Insights in Google Earth: New Earth AI models are now powering Gemini capabilities in Google Earth. Users can instantly find objects and discover patterns from satellite imagery, such as spotting a dried-up river to predict the risk of dust storms or monitoring harmful algae blooms in water supplies. 🌍 Bringing Earth AI to Google Cloud: We are making Earth AI Imagery, Population, and Environment models directly available to Trusted Testers on Google Cloud. This allows businesses to use their own proprietary data alongside our models to tackle challenges from environmental monitoring to disaster response. 🌍 Real-World Impact: Partners like the WHO AFRO are using Earth AI to predict areas at risk for cholera outbreaks, and McGill and Partners are using it to predict property damage before a storm to help clients pay claims faster. We are committed to supporting organizations building solutions in public health, crisis response, and resilience. Dive deeper into the research behind the new Earth AI innovations →https://goo.gle/47kTHyN Express interest in getting early access →https://goo.gle/4nZE9HP Read more →https://goo.gle/3WTe0ym
-
-
Google recently released Nano Banana, the latest Gemini native multimodal generation model. Check out a demo today at 3:30pm HST at the #ICCV2025 Google booth, highlighting some of its capabilities, such as image editing, interactive generation, & more. https://lnkd.in/d8V9-J73
-
Stop by the #ICCV2025 Google both at 2:30pm HST for a behind the scenes look at how @GoogleDeepMind's AI research transformed the movie the Wizard of Oz to make an immersive 16K experience for the Vegas Sphere. Blog: https://lnkd.in/gmbCcmFG Video: https://lnkd.in/gn7N_NYi
-
-
Building quantum computers means solving a core dilemma: how to control qubits without them losing information. Hear from Yu Chen, Director, Quantum Processor, about the "secret sauce" behind our Willow chip. It enabled Quantum Echoes to run 13,000x faster than a supercomputer, a verifiable quantum advantage. For a full technical deep-dive on the Willow hardware behind our breakthrough algorithm, Quantum Echoes, read the blog → https://goo.gle/47C7jH9