To address the growing challenges posed by geohazards and their cascading impacts, we invite contributions that advance reliable and scalable ML/AI approaches for the detection, monitoring, and prediction of geohazards (e.g., earthquakes, tsunamis, volcanoes, landslides, and surface subsidence). We especially welcome works that:

1. 1leverage multimodal and generative geospatial foundation models, including agentic and multi-agent approaches;
2. 2integrate multi-scale and multimodal sensing (e.g., optical/SAR, InSAR, distributed fiber-optic sensing) with ML/AI for pre- and post-event assessment of geohazards and their cascading impacts across diverse settings (urban, remote, post-disturbance, and cold-region environments);
3. 3develop interpretable and knowledge-guided ML/AI to reveal driving factors and physical mechanisms;
4. 4evaluate and improve model robustness in extreme and data-scarce scenarios through cross-region/cross-scenario transfer, uncertainty quantification, real-time data fusion, and parameter-efficient tuning; and
5. 5develop open-source, multi-scale, AI-ready geohazard datasets and benchmarks to support the pre-training, adaptation, and evaluation of next-generation geospatial foundation models for real-world deployment.