Our new thermal resilience tool uses AI-enhanced satellite and aerial imagery to help cities cope with dangerously high temperatures (and hopefully reduce the resulting health effects). The tool combines AI-based object detection with other models to analyze factors such as the amount of green space a city has or the ability of building roofs to reflect sunlight. This helps city planners and local governments determine which cooling interventions, like tree planting, can have the most impact on temperatures, down to the neighborhood level. To begin, we’re testing the tool in 14 U.S. cities, where officials are using it to determine which neighborhoods are most at risk from extreme heat and develop a relief plan. As temperatures rise globally, especially for people living in urban heat islands, AI can be part of the solution.
Google Research partners with leaders in the fire community to build FireSat, a constellation of satellites specifically designed to detect and track wildfires as small as a classroom (approximately 5 x 5 meters). Using FireSat, authorities will be able to access high-resolution images updated globally every 20 minutes, helping them respond to fires as quickly as possible. This will solve an important problem in firefighting: until now, firefighters had to use satellite images that were either low-resolution or updated only a few times a day, making it difficult to detect fires until until they become bigger than a football field. Once initiated, emergency responders will receive this critical information when fires are potentially easier to contain.
We created our basic bioacoustic model Acoustic representations of healthor HeAR, available to researchers. HeAR is designed to help researchers create models that can listen to human sounds, such as coughing, and detect early signs of illness. This technology has the potential to help detect diseases like tuberculosis. TB is treatable, but every year millions of cases go undiagnosed – often because people do not have easy access to health services, particularly in countries without sufficient medical resources. This type of AI can help. For example, a respiratory care company based in India is studying how HeAR could improve its early detection of tuberculosis based on cough sounds.