AI for Good webinar: AI for Earth and Climate - From understanding processes to effective action

AI for Earth and Climate: From satellite pixels to causal insights for effective action (speaker Gustau Camps Valls)
Artificial Intelligence is rapidly transforming Earth and climate sciences, empowering us to understand, monitor, and anticipate complex environmental and societal phenomena at unprecedented scales. In this talk, we explore the role of AI across multiple domains: from mapping crop health and ocean biogeochemistry to estimating air quality and predicting extreme weather events. We show how AI models are used not only for detection and prediction but also for interpretation, thanks to advances in explainable and causal machine learning. As climate risks intensify, AI enables actionable insights, from optimizing urban sustainability to understanding drivers of human displacement and evaluating the impacts of humanitarian interventions. Through examples that combine satellite data, socioeconomic indicators, and causal discovery, we demonstrate how AI contributes to evidence-based policymaking and strengthens our collective capacity to address critical global challenges. This talk argues that the future of climate resilience and sustainable development lies in integrating AI, Earth observation, and domain knowledge, to not only predict outcomes, but to understand and shape them.

Learning the Earth: From understanding processes to early warning of climate extreme impacts with AI (speaker: Markus Reichstein)
Artificial intelligence is transforming how we study and respond to the Earth system. This talk explores how machine learning, when thoughtfully integrated with domain knowledge, can both uncover hidden patterns in complex climate data and help anticipate emerging risks. Beginning with advances in interpretable deep learning, we show how AI can reveal meaningful process representations from data, bridging the gap between black-box models and physical understanding. Building on this foundation, we explore how integrated AI frameworks—combining Earth observations, causal inference, and expert knowledge—can provide early warnings of impacts of climate extremes with a focus on ecosystems and their services. These approaches mark a shift from reactive to anticipatory risk management and highlight the potential of AI not only for scientific insight but for supporting decisions in a changing climate.