The European Union's Common Agricultural Policy (CAP) represents a significant portion of the EU budget, but its subsidy distribution system is vulnerable to fraud. The 2025 OPEKEPE scandal in Greece, involving fraudulent claims for agricultural and livestock subsidies, highlighted the urgent need for robust, automated verification systems. This paper presents Green Horizon, a multi-modal AI system designed for large-scale agricultural monitoring and subsidy compliance verification. The system integrates real-time Copernicus Sentinel-2 satellite data with national land registry (Cadastre) and agricultural payment agency (OPEKEPE) databases. Its core, the "AI Pantheon" engine, uses a time-series analysis of spectral indices like NDVI and NDMI to perform crop classification, monitor crop health, and detect discrepancies between declared and actual land use with over 95% accuracy. We describe the system's architecture, its successful demonstration to the Region of Central Macedonia and the Greek Ministry of Agriculture, and its subsequent submission to the European Union. Furthermore, we propose a novel extension for automated livestock counting using high-resolution satellite or drone imagery and deep learning, directly addressing the vulnerabilities exposed by recent fraud cases. Green Horizon offers a scalable, data-driven solution
Grigorios Tsinaforniotis (Sun,) studied this question.