Ethical and Responsible Artificial Intelligence for Earth Sciences

The use of Artificial Intelligence (AI) in the Earth sciences has grown exponentially over the past few years. With the increasing role of AI in a wide-ranging set of Earth science applications such as weather prediction, it is crucial that AI systems are developed in an ethical and responsible manner. This project expands on earlier work by the research team to identify, measure, and address multiple ways that AI could be inadvertently misused within environmental and Earth sciences.

Through workshops, research with public and private partners, and by developing teaching materials on ethical and responsible AI for the Earth sciences, this project will contribute to generating more trustworthy AI along with upholding national safety and welfare.

This project advances understanding of trustworthy AI by co-producing research with public and private partners on ethical and responsible AI in the Earth sciences. The research team aims to learn how bias from a variety of sources can affect AI for Earth sciences and can be mitigated through five main activities. Data-related challenges specific to the Earth sciences make existing approaches for bias identification and measurement inapplicable in weather and climate.

In its first activity, the team will extend its categorization of biases for AI in the Earth sciences, to enable AI developers to identify, measure, and mitigate serious biases in their applications. Second, informed by this work on measuring and mitigating bias, the team will then co-develop a bias assessment framework for AI for weather and climate, with research partners across public and private sectors.

Third, while seeking to work with government agencies and other partners, the team will then develop and test a risk management framework focused specifically on AI for the Earth sciences. Building on this, the team will investigate continued co-development and broader adoption of this AI risk management framework by government agencies. Finally, the team will co-develop teaching materials in ethical and responsible AI for weather and climate in collaboration with other AI ethics researchers, with an emphasis on co-producing risk assessments with community partners potentially affected by AI.

This project is jointly funded through the ER2 program by the Directorate for Social, Behavioral and Economic Sciences, the Directorate for Biological Sciences, and the Directorate for Computer and Information Science and Engineering.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.