Recommendations

Addressing challenges in conducting AI/ML research

Data science and AI have incredible potential for applications in large-scale data analysis to understand the natural environment from various angles, such as for biodiversity monitoring, detecting icebergs in vast polar oceans and tracking wildlife from space. However, not everyone involved in environmental research and data analysis has the technical expertise necessary to use the latest available tools. Furthermore, researchers and practitioners often lack a practical understanding of best practices in data science and lack access to an interdisciplinary network of experts essential for addressing global challenges affecting different parts of our society. These, in addition to broader challenges presented by climate change, add to technical challenges associated with the applications of data science and AI.

Categorised under these four overarching objectives, in this paper, we describe some of the most important challenge areas related to building, testing and implementing AI/ML solutions for monitoring, forecasting, simulating and responding to climate change.

• (Facilitate) Interdisciplinary and Inclusive Collaboration
• (Create) Open, Reproducible and Ethical AI Technology
• (Pioneer) Cross-Disciplinary Data Integration and Real-World Deployment
• Informing community of ethical use of AI