Artificial intelligence and the labor market in Spain: Occupational exposure and estimated effects on employment
AI disruption
Fecha: mayo 2026
Francisco Rodríguez Fernández
SEFO, Spanish and International Economic & Financial Outlook, V. 15 N.º 3 (May 2026)
The relationship between artificial intelligence and employment has shifted substantially since Frey and Osborne′s (2017) estimate that 47% of U.S. jobs faced high automation risk, a figure now subject to methodological revision. Second-generation indices, built on task-level analysis rather than occupational categories, find that exposure to generative AI is concentrated among educated, higher-wage workers, not the routine manual jobs of the previous paradigm. Experimental evidence reinforces this picture: productivity gains are significant, but benefit less experienced workers the most, while aggregate TFP gains over a ten-year horizon are estimated at under 1%. Applying an adapted version of the AI Occupation Exposure index to Spain′s CNO-11 occupational classification, the model projects gross job displacement of between 1.7 and 2.3 million positions over 2025–2035, with a central estimate of 2.0 million concentrated in administrative, technical support, and scientific professional roles. Against this, complementarity effects benefit an estimated 3.1 million workers in services and manufacturing, and new occupation creation is projected at 1.61 million, placing the net loss at around 400,000 jobs in the baseline scenario. The distribution of that loss across geography, sector, and educational level is highly uneven, and the transition window depends critically on the pace at which displaced workers can access reskilling. Spain′s position at a historic employment peak as AI adoption accelerates represents an opportunity to manage that transition from a position of relative strength.