Descriptive Summary:

Automation, Artificial Intelligence, and Employment in the European Union: An Early-Stage Meta-Analysis of Macroeconomic Evidence

This working paper provides the first EU-focused meta-analysis of macroeconomic, causal studies examining the employment effects of automation and artificial intelligence. Synthesizing 15 quantitative studies and selected grey literature, it finds a modest net positive employment effect (approximately 1.5-2%), but with pronounced sectoral, regional, and skill-based disparities.

Manufacturing-heavy and less technologically advanced regions face greater displacement risks, particularly among routine and low-skilled workers. In contrast, knowledge-intensive and digitally advanced economies benefit from productivity gains and high-skilled job growth. The analysis confirms patterns consistent with skill-biased technological change and task-based complementarity.

A central finding is the scarcity of robust macro-level causal evidence specifically on AI, distinct from broader automation. The paper also introduces and validates an AI-assisted methodology for systematic evidence synthesis, demonstrating high reliability in human-AI screening collaboration.

The findings underline the need for targeted reskilling, regional policy differentiation, and stronger empirical research on AI-specific labour market effects.