COAST-AId: a custom large language model application supporting multi-hazard risk assessments in the Veneto region

The global climate is undergoing an unprecedented rise in temperature, contributing to the increased frequency and intensity of extreme events worldwide. Coastal areas, recognized as critical hotspots of climate change, face amplified vulnerabilities due to their dense population, interconnected economic activities, and delicate ecosystems. These regions are particularly threatened by sea-level rise and more frequent extreme weather events, underscoring the urgent need for innovative and comprehensive strategies to enhance climate resilience and safeguard their future. Addressing these challenges requires a paradigm shift toward integrated, multi-hazard, and multi-risk approaches able to capture the intricate interplay of overlapping risks.
In this context, we present COAST-AId, a custom Large Language Model application designed to support multi-risk assessment and adaptation planning in the Veneto coastal region. COAST-AId facilitates the application of the climate risk assessment framework, outlined in the first European Climate Risk Assessment (EUCRA) report, focusing on its key components such as risk identification, analysis, and policy evaluation. By employing state-of-the-art prompt engineering techniques, COAST-AId has demonstrated its capability to generate relevant outputs for the case study, systematically evaluated using specific assessment metrics. This tool prioritizes risks for the Veneto coastal areas, providing essential insights to guide the development of Disaster Risk Management (DRM) pathways offering valuable support for refining the Veneto region Strategy for Climate Change Adaptation, with a particular emphasis on coastal challenges.
The development and application of COAST-AId were deeply integrated into the objectives of the MYRIAD-EU project, fostering the collaborative engagement with local stakeholders and experts. This participatory process played a central role in evaluating the tool’s performance, identifying critical vulnerabilities, and uncovering opportunities to enhance risk reduction and adaptation strategies. The findings highlight the transformative potential of AI-driven technologies in advancing the understanding of multi-risk dynamics, optimising decision-making processes, and enhancing resilience-building efforts in coastal areas.