Navigating the Legal Architecture of AI
Abstract
A state that ignores a 95.6% accurate landslide prediction and then invokes force majeure is not applying a legal doctrine it is abusing one. Force majeure and the Act of God defence share a single operative requirement. Machine learning has destroyed that requirement. AUC values for flood and landslide prediction routinely exceed 0.90 rendering catastrophic events institutionally foreseeable before they occur. The EU AI Act (Regulation (EU) 2024/1689) reframes the analysis. Classifying disaster-management systems as high-risk AI raises the standard of care for public authorities and makes force majeure progressively harder to sustain as a complete defence. Explainability is a due-process obligation here, not a technical amenity: communities subject to AI-driven evacuation orders hold a legitimate interest in understanding the basis for those decisions. The concern that strict liability will deter disaster-AI deployment is real and deserves a substantive answer. This chapter proposed a standards of domestic enforcement, pre-deployment sandboxes, safe harbours, and compensation funds.
