AI-driven Earth Observation for monitoring archaeological looting: advances, opportunities, and future directions

Illicit excavations of archaeological sites constitute a major threat to cultural heritage, causing irreversible damage and information loss, an issue persisted for several years. However, at the beginning of the 21st century, advances in satellite Earth Observation (EO) and Artificial Intelligence (AI) are transforming how such activities can be detected, monitored, and understood. High-revisit optical constellations (e.g., PlanetScope, Sentinel-2) and synthetic aperture radar (SAR) missions (e.g., Sentinel-1) now provide dense, multi-temporal data for near-continuous surveillance. In parallel, AI-based methods - including deep change detection, temporal modeling, and multimodal fusion - enhance sensitivity to small, transient signals and enable scalable, reproducible monitoring. This paper synthesizes emerging approaches that integrate EO driven by AI for identifying and analyzing archaeological looting, discusses their technical implications, and highlights future directions for operational, transparent, and globally accessible heritage monitoring systems.