Building AI to accelerate science and improve lives

Google published an overview of its AI-for-science and society work, authored by James Manyika, SVP of Research, Labs, Technology & Society. The company said its technologies now support more than 300 languages spoken by 7 billion people, representing 86% of the global population, and that it released interactive insights through its AI & Economy ATLAS, described as its most comprehensive look at how people use AI globally.

On health, Google said AlphaGenome Atlas maps the predicted impact of all 9 billion possible single-letter genetic changes across the human genome and is openly available to researchers. It cited a breast cancer study with Imperial College London and the U.K. NHS showing AI detected 25% of interval cancers previously missed across mammograms of 175,000 women. Its chest X-ray tool for tuberculosis, used by Nexus Intelligence, has screened over 25,000 X-rays in 40 locations across six nations; Google noted roughly 40% of infected people worldwide go undiagnosed. A diabetic retinopathy model has supported more than 1.15 million screenings, with plans to reach 6 million over the next decade. AlphaFold, which predicted all 200 million known protein structures, is used by 4 million researchers in 190 countries.

For weather and disasters, Google introduced WeatherNext 3, which it says delivers 50% more accurate precipitation forecasts a day or more ahead and combines satellite observations with AI to produce hourly forecasts without heavy supercomputing. Flood Hub forecasts now cover 2 billion people in more than 150 countries. In 2025, monsoon predictions informed 38 million farmers in India, and more than 520 crisis alerts on Google Search reached over 75 million users. The Earth AI Planetary Prediction Engine, an autonomous system driven by plain-language instructions, pinpointed 83% of emerging Ebola hotspots ahead of time in the Democratic Republic of the Congo and doubled local food security forecasting accuracy in Nigeria.

On learning and work, Google said it partnered with ISTE+ASCD to provide AI literacy training to all 6 million U.S. educators, and has provided more than $1 billion globally in training and skilling initiatives, helping over 100 million people. It said one language may have 100 dialects, and it open-sourced speech and image datasets for 109 Indic languages with IISC-Bangalore and ARTPARK. Google also said it has given more than $1 billion across 1,700+ research institutions since 2006, and framed AI's economic benefits as neither automatic nor guaranteed, expecting jobs to decline, grow and change.

Why it matters: Developers and researchers gain open access to resources such as AlphaGenome Atlas and datasets for African and Indic languages, while several models, including WeatherNext 3, are already deployed in Google products and public alerting systems.