Data-Driven Insights on Human Security Integration in EU Green Deal Policies

This paper proposes a novel, data-driven methodology to systematically assess how European Green Deal policy texts address various Human Security Aspects, including newly acknowledged technological vulnerabilities. By analyzing official EU documents using advanced semantic modeling and transformerbased embedding techniques, we demonstrate how machine learning can identify thematic alignments or gaps in addressing human security within policies explicitly connected to the Sustainable Development Goals. Our approach, which employs Sentence-BERT models and cosine similarity measures, reveals that while EU Green Deal policies integrate all eight human security dimensions with relatively balanced coverage, economic, food, and community security receive slightly more emphasis than personal, political, and technological security aspects. These findings illuminate both strengths and opportunities for enhancement in current policy discourse, aiding stakeholders in designing interventions that explicitly integrate comprehensive human security perspectives.