Structural Risk Under Automation: Measuring Labor Market Risk Across EU Regions

This thesis develops the Composite AI Labor Market Risk Index (CAILM-RI) to evaluate structural labor market risk across 240 European NUTS-2 regions in the context of accelerating artificial intelligence (AI) and digital transformation. Rather than predicting employment outcomes, the index aims to diagnose latent vulnerabilities and adaptive capacities embedded in regional labor markets between 2013 and 2023. The framework draws on 15 harmonized indicators grouped into three core dimensions: technological exposure, socioeconomic vulnerability, and resilience capacity. Indicators are normalized using fixed bounds and aggregated using the Adjusted Mazziotta–Pareto Index (AMPI), a non-compensatory method that penalizes internal imbalances, ensuring that structurally fragmented profiles are not misclassified as resilient. The results reveal substantial territorial disparities in composite scores. Several economically advanced regions exhibit elevated risk due to asymmetries between innovation intensity and social or institutional buffering capacity (Rodríguez-Pose, 2018). In contrast, certain mid-performing regions demonstrate more coherent structural profiles, with better alignment between labor market inclusion, education systems, and policy mechanisms. A four-part typology, Consistently Resilient, Improving, Declining, and Persistently Vulnerable, is introduced to capture the level and stability of regional risk profiles. Although the index does not claim causal inference or forecasting capacity, it provides a replicable, transparent, and policy-relevant framework for identifying systemic labor market fragilities in the age of AI. By focusing on multidimensional structural alignment, the CAILM-RI contributes to more territorially sensitive approaches to employment resilience, complementing existing EU and national labor policy instruments (OECD, 2021; Cirillo & Guarascio, 2024).