The new automation divide: Understanding the role of AI in employment shifts, distribution, and future policy solutions

This paper examines the labour-market implications of artificial intelligence (AI) focusing on employment, distribution, and economic governance within the European Union. It presents the reader with the complicated literature around automation technologies, and their impacts on labour. Arguing that AI represents a qualitatively new phase of automation but continues to demonstrate similar trends to previous waves, only this time targeting routine cognitive and white-collar jobs. The paper situates AI adoption within existing theory, demonstrating how power relations and firm-level incentives shape whether technology substitutes or augments labour.

Drawing on the task- and skill-based framework of the automation literature, the study combines a critical review of debates with an empirical analysis of Eurostat data from 2015 to 2024. The findings indicate that employment growth in the EU has been increasingly concentrated in non-routine cognitive occupations, while routine clerical and administrative roles have experienced marked decline. These long-term trends have, however, been accelerated since the diffusion of generative AI in 2022, especially in exposed sectors. Additionally, asymmetric AI adoption across firm sizes, with large enterprises leading deployment and disproportionately affecting a large pool of workers that are concentrated in such large firms.

The paper concludes that AI-driven productivity gains are unlikely to automatically be shared without proactive economic governance. Active labour-market policies, while necessary, are insufficient to address the effects of AI. The study therefore explores the feasibility of AI-related taxation as a complementary policy instrument to discourage excessive labour substitution, restore fiscal capacity, and incentivise labour-augmenting forms of AI deployment. EU AI governance must explicitly protect entry-level employment, skill formation, and labour’s share of income if technological change is to be aligned with social objectives.