Monitoring Sustainable Development Goals in European Legislation using Hybrid AI

There is an urgency to detect in legal acts the corresponding provisions where a policy is implemented, to track its evolution over time, to measure the effectiveness of the norms, and to evaluate the impact on society. From this perspective, the Sustainable Development Goals program (SDG) provides a fundamental instrument for monitoring ground basis pillar of the world wide policies. In this work, we propose a method which leverages both the structural nature of legislative documents in AKN-XML and unsupervised machine learning to perform a match between individual articles and definitions and the 2030 Agenda for Sustainable Development Goals, yielding a more fine-grained annotation of individual articles and definitions, instead of the preexisting document-level annotation. Our work provides better traceability of the SDGs policies in the EU legislation permitting the legislator to detect the articles where the association is weakest. During the legal drafting, our tool could be integrated into the editor to suggest better legal definitions for improving the implementation of the SDGs