THE IAEA’S INNOVATIVE APPROACH TO ADDRESS THE CHALLENGES IN THE COLLECTION AND ANALYSIS OF SAFEGUARDS RELEVANT INFORMATION

The IAEA’s Division of Information Management (SGIM) is responsible for processing and analysing all three sets of Safeguards relevant information (State-declared; information from infield activities; and all other safeguards relevant information). This paper outlines the work of SGIM and describes the current and future challenges for information collection and analysis. In addition, it explains the innovative machine learning (ML) and artificial intelligence (AI) approaches being developed by SGIM. Examples include the assessment of relevance and classification of open source information, satellite imagery change detection, and automated comparison of particles from environmental samples against a global sample database. Additionally, AI/ML models are being used to provide realistic simulated nuclear material accounting (NMA) and verification data to external contributors for developing statistical evaluation methodologies or internal training. These approaches have shown to enhance the process of collection and review of Safeguards relevant information and contribute to the overall effectiveness and efficiency of the analytical process.