CrisiSense-RAG: Crisis Sensing Multimodal Retrieval-Augmented Generation for Rapid Disaster Impact Assessment

Researchers unveiled CrisiSense-RAG, a multimodal AI that fuses real-time social media reports with delayed satellite imagery to assess disaster impact. It solves a critical timing mismatch where post-event images often underestimate peak flooding. In zero-shot tests on Hurricane Harvey across 207 areas, it reduced flood extent errors to 10.94-28.40% and damage severity errors to 16.47-21.65%. The system prioritizes social evidence for flood extent and imagery for structural damage without disaster-specific training.