European Central Bank sees AI’s Impact on US employment as limited for now

The effects of artificial intelligence (AI) on overall employment appear to be limited so far, although there is increasing evidence that it is negatively affecting employment in specific subgroups, particularly young workers in highly exposed occupations, according to an article published by the European Central Bank (ECB), which focuses its analysis on the United States, where such effects are likely to have become visible earlier than in other major economies.

The authors of the study acknowledge a significant potential for AI to transform labor markets, emphasizing that its impact on employment growth can be both positive and negative, by enabling greater productivity but also displacing jobs, so its net impact depends on the relative importance of these effects in each country.

In this regard, the widespread adoption of AI in US companies would generate positive effects on employment, although these conceal considerable heterogeneity among different occupational groups, while initial evidence in the European Union suggests that companies adopting AI technologies experience greater productivity increases, without the technology replacing labor in the short term, in line with a recent ECB survey indicating that companies with high levels of AI adoption or related investment are more likely to hire additional staff.

Specifically, the study notes an average 4% decrease in employment in positions with a high risk of replacement by AI, such as economists or graphic designers, between 2019 and 2025, while in positions with low risk of replacement, such as electricians or high school teachers, it increased by 13% during the same period.

As a result, the composition of employment in the US has changed, and the proportion of low-risk jobs in the total US labor market has risen from 23% to 25%, while the proportion of high-risk jobs has decreased from 35% to 33%.

"AI has already caused a reallocation of jobs in the US labor market," the authors note, for whom the data indicates a growing gap between employment growth in occupations with a high risk of replacement by AI and occupations with a low risk of replacement, which is consistent with studies showing that AI is affecting employment growth in specific occupational subgroups.

Thus, while the consequences of AI for aggregate employment to date remain inconclusive, the analysis reveals that it has had a relative impact on employment growth in the US since 2019, which has accelerated since the launch of ChatGPT at the end of 2022.

On the other hand, this relative impact of AI on job creation has not yet translated into significant differences in wage growth, although the authors admit that empirical evidence on this is scarce, noting that over time, as the labor market continues to adjust and AI tools become more productive, the effects on income could be more pronounced.