5.2 How the NWS is Training Artificial Intelligence to Address Language Accessibility Gaps in Weather Information

Tuesday, 30 January 2024: 8:45 AM
Johnson AB (Hilton Baltimore Inner Harbor)
Monica L Bozeman, NWS, Silver Spring, MD; and A. Montanez, C. Rohrbach, J. E. Calkins, R. Chai, B. Johnson, M. Torres, A. Reynes, I. C. Colon-Pagan, R. Vasquez, E. Rodriguez, F. E. Morrone, N. Vaz, M. Bailey, A. L. Lago, K. Farina, C. Moore, L. E. Llewellyn, and J. E. Trujillo-Falcón

In order for the National Weather Service (NWS) to achieve its strategic vision of a Weather-Ready Nation, the agency needs to address communication gaps to better reach people with Limited English Proficiency (LEP). This communication gap puts approximately 25 million U.S. LEP people in potential danger during severe weather because they do not understand NWS messaging. While limited outreach material is available in Spanish, the NWS remains unable to effectively communicate hazard and weather information in real time to communities with LEP due to the fact that nearly all NWS products are only distributed in English.

Since the NWS has limited linguistic abilities beyond English, it has become pivotal to leverage Artificial Intelligence (AI) to instantaneously translate NWS Watches and Warnings to vulnerable multilingual communities. There are unique challenges associated with translating into multiple languages as it is difficult to apply a one size fits all approach for every target language. Therefore, the AWIPS Program has developed the NWS Automated Language Translation Project. The goal of the project is to set the groundwork for developing a nationally-supported, consistent, automated approach to translate English NWS products at scale across multiple languages, Weather Forecast Offices (WFOs) and National Centers.
This presentation will provide an overview of the AI translation technology used by the NWS, recent successes from our pilot projects, lessons learned, and highlight several linguistic and cultural challenges involved in producing high quality and understandable translations faced by the NWS Automated Language Translation Team. In addition, the talk will provide demographic insights across WFO regions to better understand the communities they serve, and highlight a new website developed to integrate and display translated weather information into local LEP communities. Taken together, these efforts will contribute to making the nation more Weather Ready regardless of what language is spoken.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner