When a tornado ripped through Saragosa, Texas in 1987, a Spanish-language radio station broadcast the National Weather Service warning—but the words fell flat. The English term "warning" had no equivalent in Spanish, and 151 of the town's 183 residents were killed or injured. Nearly four decades later, that tragic gap in communication has finally begun to close.
The National Weather Service is now deploying artificial intelligence to translate life-saving weather forecasts and alerts into Spanish, Chinese, Vietnamese, Samoan, French and other languages, reaching the 68.8 million people across the United States who speak a language other than English at home. What was once a painstaking manual task—consuming up to an hour per translation product for bilingual forecasters already stretched thin with operational duties—now takes 5 to 7 minutes, with accuracy scores above 95 percent.
The breakthrough came through a partnership between the NWS and LILT, an AI translation platform whose patented training process allows large language models to learn the specialized terminology and messaging conventions of meteorology. University of Illinois Urbana-Champaign professor Joseph Trujillo-Falcón, who leads the ALERTAS lab, and NWS scientist Monica Bozeman documented the program in a peer-reviewed study published in the journal Artificial Intelligence for the Earth Systems.
"Translating weather forecasts has always been a critical, time-consuming task, often added to the plates of bilingual forecasters managing full operational responsibilities," Trujillo-Falcón explained. The scale of this work is now evident: more than 30 NWS offices across the country have begun using the system, and starting with the 2025 hurricane season, the National Hurricane Center officially began issuing AI-translated Spanish advisories for all hurricanes affecting the Pacific and Atlantic coasts—providing lifesaving information not only to communities within the U.S. but also across Latin America.
The geographic scope of the initiative reflects careful research. Graduate student Liam Llewellyn overlaid geographic information systems, weather, and census data to identify which language translations are needed for all 122 NWS offices nationwide. This multidisciplinary approach—combining meteorology, technology, communication studies, and geography—ensures that the program reaches communities most likely to need it.
Beyond emergency response, the implications ripple outward. Trujillo-Falcón noted that tourists from around the world who visit the United States can now receive life-saving weather alerts in their native languages, benefiting both their safety and local economies. As the world prepares for the World Cup coming to American stadiums, the NWS has made an agency-wide effort to provide decision-support services in multiple languages in the event of dangerous storms.
The team is already exploring the next frontier: how the public actually responds to AI-translated weather information. Collaborations with Spanish and Portuguese professors and researchers from the University of North Dakota, Pace University, Colorado State University, and the University of Oklahoma will help evaluate the social dimensions of this technology. For communities like Saragosa—where a single mistranslation cost lives—this moment represents not just technological progress, but a recognition that life-saving information has no language barrier.