ECB: AI boom has limited impact on US jobs, study finds

The European Central Bank released a study Monday that examined the labour-market effects of rising artificial intelligence adoption in the United States. The analysis covers 2019 through 2025 and focuses on occupational changes tied to AI exposure.

The study uses an occupation-level index that groups jobs into low, medium and high AI substitution risk. Researchers compared employment growth across these groups using 2019 as a base year and controlled for sector-specific trends, the economic effects of the COVID-19 shock and other unobserved industry differences to isolate the impact of AI exposure.

At the aggregate level, the report found no broad decline in total U.S. employment or wages attributable to AI through 2025. The study cites recent research and ECB survey data indicating that companies with higher AI adoption or AI-related investment were more likely to hire additional staff. Early evidence from the European Union cited in the report points to higher productivity at firms adopting AI without significant short-term labour replacement.

The study identified a clear reallocation within the U.S. labour market. Employment in occupations classified as high risk of AI substitution fell by more than 4% on average between 2019 and 2025. Examples of roles the report lists as highly exposed include economists and graphic designers. By contrast, employment in low-risk occupations such as electricians and high school teachers rose by about 13% over the same period.

The share of low-risk jobs in total U.S. employment increased to 25% from 23%, while the share of high-risk jobs declined to 33% from 35%. When comparing growth across risk categories, the report found that jobs with a high risk of AI substitution expanded roughly 15 percentage points less than low-risk jobs between 2019 and 2025, after accounting for other factors.

The analysis notes that the divergence in occupational outcomes has accelerated since the public release of large-language-model applications in late 2022, suggesting labour-market effects may become more visible as AI adoption spreads across industries.

The report frames AI’s labour-market influence as the result of two opposing forces: the technology can raise productivity and enable firms to expand and hire, while it can also replace workers in tasks that can be automated. The study documents measurable shifts in job composition across occupations in the United States through 2025 but reports limited aggregate effects on overall employment and wages to date.