Developing trustworthy AI for weather and climate

Amy McGovern,Philippe Tissot,Ann Bostrom

DOI: https://doi.org/10.1063/pt.3.5379

IF: 3.5

2024-01-01

Physics Today

Abstract:By improving the prediction, understanding, and communication of powerful events in the atmosphere and ocean, artificial intelligence can revolutionize how communities respond to climate change.

physics, multidisciplinary

What problem does this paper attempt to address?

This paper discusses how to harness artificial intelligence (AI) technology to enhance the prediction, understanding, and communication of weather and climate events, thus innovating community response to climate change. With the changing climate, the impact of extreme weather on global species and ecosystems is intensifying, and advancements in AI hold promise for providing earlier and more accurate warnings, enabling people to mitigate the effects of these events. The paper highlights that while AI has been applied in weather and climate prediction, its practical adoption is still limited. To establish trustworthy AI, interdisciplinary collaboration is needed to ensure that AI methods align with user needs and provide reliable forecasts.

The paper proposes a comprehensive approach to developing trustworthy AI, including co-development with end-users, risk communication, environmental science, and ethical and responsible AI. Through explanatory AI methods, users are assisted in understanding the workings of AI models, increasing trust. The paper also showcases the application of AI in predicting and responding to cold wave events, using the example of a sea turtle rescue program, and illustrates how trust can be established through close collaboration with users. Furthermore, the paper discusses the potential of AI in predicting other extreme weather events such as thunderstorms, tornadoes, and hail, emphasizing the importance of avoiding bias and ensuring ethical responsibility in AI development.