A copula-based human–artificial intelligence collaborative decision-making approach for multi-hazard risk assessment

Accurately quantifying multi-hazard coupling risks is essential for developing effective emergency management strategies. This study proposes an integrated decision-making framework that synergizes human intelligence (HI) and artificial intelligence (AI) to address the challenges of nonlinear risk assessment. Leveraging multivariate dependency modeling and guided by explainable AI (XAI), the framework is empirically validated in the Greater Bay Area (GBA), a region prone to compound hazards. Results show that HI-AI collaboration substantially enhances the model's ability to capture intricate hazard interdependencies, improves AUC by 9.41% through the integration of expert-derived rules, and benefits greatly from diverse expert meta-knowledge. This approach offers a robust and interpretable tool for managing multi-hazard risks and improving decision-making in complex emergency scenarios.