As data-driven models increasingly rival or complement physics-based systems, a central question remains open: can AI reliably forecast the events that matter the most, the extremes that drive real-world impacts, and not just the average state of the atmosphere? This session brings together the latest advances in machine learning (ML) and artificial intelligence (AI) for forecasting weather, projecting climate, and simulating extreme events.

We invite contributions spanning the full range of timescales and methods, including but not limited to:
*data-driven and foundation weather models for short- and medium-range forecasting;
*generative and probabilistic approaches (e.g. diffusion models) for forecasting, downscaling, and uncertainty quantification;
*ML for sub-seasonal to seasonal (S2S) prediction and longer-term climate projections;
*hybrid AI-physics approaches that embed physical constraints into data-driven models or improve the representation of climate variables in numerical models and datasets;
*detection, attribution, and anticipation of extreme events such as hurricanes, floods, heatwaves, droughts, and compound extremes.

We particularly encourage submissions that go beyond forecast skill to address impacts on infrastructure, ecosystems, health, or energy systems, and that engage with questions of trust, explainability, and generalization to unseen or out-of-distribution extremes.

By bringing together experts from AI, data science, meteorology, climate science, and impact modelling, this session aims to foster interdisciplinary collaboration and push the boundaries of AI-driven understanding and prediction of extreme weather and climate events. We warmly welcome submissions from early-career scientists, established researchers, and industry professionals alike.