AI in the Practice of Geoscience

Beyond the classroom, geoscience as a discipline has been and will continue to be impacted by AI advances. In preparing future geoscientists, we need to track these changes so that the classroom experience aligns with those needs.

Disciplinary Norms around Generative AI Use

Generative AI can be used to support a variety of tasks geoscientists engage in. From ideation and literature searches, to paper and grant writing, users are experimenting with what AI can do and what norms we should establish as a discipline. Should journals prohibit the use of AI in preparing papers for publications, or is AI a key tool for allowing full participation in scientific exchange for non-native English speakers? How much AI 'help' can you get before having to declare its involvement in your papers or grant proposals, and what does that declaration look like? As these norms evolve, we'll want to track them and share them with our students. Here are some starting points:

• Effective and Responsible Use of AI in Research - a guide for graduate students that touches on a range of relevant issues shared by Georgia Tech
• The American Geophysical Union (AGU) refers authors to Wiley's Best Practice Guidelines on Research Integrity and Publishing Ethics
• The Geological Society of America (GSA) refers authors to the COPE Guidelines on Artificial intelligence and authorship

 

How are you navigating the evolving disciplinary norms around AI use?

Machine Learning and AI as a Tool for Geoscience Research

Beyond the highly publicized world of large generative AI systems, there are a wide range of AI tools, especially in the field of machine learning that are widely used in geoscience research. As with any advance in the field, we need to reflect on how our teaching can help students get up to speed with these approaches.

• A Two-Step Approach to Training Earth Scientists in AI - an EOS article describing one effort to help geoscience researchers get up to speed with these technologies.
• The Teach the Earth collection also includes a growing number of examples of how these tools might be introduced to students, including:
  • Supervised Machine Learning Readiness Nicole Corbin and Thomas Martin, University Corporation for Atmospheric Research (UCAR)
  • Time Series Modeling and Prediction of Environmental Data M.E., M.R. Hipsey, K. Kurucz, and C.C. Carey as part of Project EDDIE
  • Constructing a data-driven rainfall-runoff model Tianfang Xu, Albert J. Valocchi, University of Illinois at Urbana-Champaign
  • Forecasting River Discharge and Runoff with a Python-based Statistical Machine Learning Tool Adnan Rajib, Analisa Gonzalez, Qianjin Zheng,  The University of Texas at Arlington

 

Have you developed teaching materials that engage student machine learning as a disciplinary tool?  Consider sharing your activities through Teach the Earth.