Fact of the Week: AI-Enabled Automation Is Positively Associated With Changes in Occupation Employment Shares
Source: Stefania Albanesi et al., “New Technologies and Jobs in Europe,” IZA Institute of Labor Economics Discussion Paper Series, no. 16227, June 2023.
Commentary: In a recent discussion paper for the Institute of Labor Economics, authors Stefania Albanesi and colleagues analyzed the AI Occupation Impact (AIOA) scores by Felten et al., exposure to AI and software scores by Webb, and employment data from the EU Labor Force Survey (EU-LFS) to examine the impact of new technologies—particularly, AI and software—on occupations’ employment shares and wages. Using the AIOA and Webb’s AI and software scores as proxies for AI-enabled automation of an occupation, the authors found that AI-enabled automation is positively associated with changes in occupation employment shares in a sample of sixteen European countries. An increase from the 25 centile to the 75 centile in the distribution of AI exposure to an occupation increased employment share for that occupation by 2.6 percent when using Webb’s scores and 4.3 percent when using the AIOA score. In other words, AI-enabled automation did not displace workers. However, evidence suggests that wages did not change when AI exposure increased.
AI exposure only affected the employment share of occupations employing specific types of workers. When the data is disaggregated by age and skill level, the authors found that AI exposure did not affect the employment share of sector-occupations employing workers with educational attainment in the lower two-thirds of the education distribution in each country. However, AI exposure is associated with a positive increase in employment share for those in the top third of the education distribution. An increase of 25 centiles in the distribution of AI exposure increased high-skill sector occupation employment share by 3.1 percent using Webb’s scores and 6.6 percent using the AIOA scores. In terms of age, employment shares of occupations with younger workers are more positively impacted by AI-enabled automation. In other words, employment increases in Europe due to AI-enabled automation are found mostly in occupations with younger and higher-skilled workers.
