The adoption of artificial intelligence tools is already leading to a reallocation of jobs within the United States labour market, according to new research published by the European Central Bank (ECB).

While the aggregate impact on total employment has remained muted, the analysis reveals a clear trend where occupations considered at high risk of substitution by AI have experienced a notable decline in job growth.

Isabella Moder, a Senior Economist at the European Central Bank, and Til Pommer examined the effects of AI on employment growth, focusing on the United States as a bellwether for global trends.

They explained that the United States is home to some of the most advanced early-adopting firms and possesses a relatively flexible labour market, making it an ideal environment to observe these shifts before they become fully visible in other major economies.

The research utilised an index developed to categorise occupations based on their susceptibility to AI replacement.

Employment in roles with a high risk of AI substitution, such as economists and graphic designers, fell by more than 4 per cent between 2019 and 2025.

Conversely, jobs categorised as having a low risk of AI substitution, including electricians and high school teachers, saw a 13 per cent increase in employment over the same period.

This divergence has altered the composition of the workforce, with the share of low-risk jobs rising from 23 per cent to 25 per cent, while the share of high-risk positions dropped from 35 per cent to 33 per cent.

When controlling for sector-specific developments and external shocks such as the pandemic, the study found a growing gap between job growth in high-risk and low-risk occupations.

Jobs with a high substitution risk grew by around 15 percentage points less than jobs with a low substitution risk between 2019 and 2025, according to the researchers.

This impact has noticeably accelerated since the launch of ChatGPT in late 2022.

Despite these shifts in employment, the researchers found that AI substitution risk has not yet translated into significant differences in wage growth.

The analysis of median hourly wage growth by occupation revealed that the risk of AI substitution has had no significant impact on wage growth since 2019.

However, the authors cautioned that this may change as the technology evolves.

“As the labour market continues to adjust and AI tools become more generative, income effects may be more pronounced,” according to the authors.

The report also contextualised these findings within the broader literature on technological change.

While the consequences of AI for aggregate employment to date remain inconclusive, the analysis finds that it has had a relative impact on US employment growth since 2019, the economists concluded.

The study aligns with recent European Union evidence suggesting that while firms adopting AI technologies often see higher productivity gains, they do not necessarily replace labour in the short term.

Instead, the immediate effect appears to be a restructuring of the workforce, favouring roles that are harder to automate while simultaneously challenging junior workers in highly exposed occupations.