Study finds AI exposure linked to lower wages, not job losses

In short: A new study finds little sign that AI is causing widespread job losses yet, but it may be holding down pay and making it harder for workers to move to better jobs.

What's going on

Official job statistics still show little clear link between how exposed a job is to AI and whether employment in that job is shrinking.

But a new paper by European researchers José Azar, Mireia Giné, and Javier Sanz-Espín looked at more than just headcounts. They combined large datasets, including millions of job ads and online professional profiles, and compared them with US government data on earnings and employment across tens of thousands of firms.

The researchers found no evidence that the most AI-exposed occupations have seen faster declines in employment than less-exposed jobs. However, they did find worse outcomes in pay and job-to-job moves. Depending on the dataset, wages for workers in the most AI-exposed roles were 5 to 10 per cent lower than for workers in the least-exposed roles.

The biggest wage drops were for the least experienced workers. The study also found that workers in AI-exposed roles were almost a third less likely to switch jobs. That matters because switching employers is often how people get meaningful pay rises, like shopping around for a better deal.

What to watch

If these patterns continue, AI may change work in a quieter way than mass layoffs. Employment can look stable while workers feel more stuck, with weaker pay growth and fewer chances to move up. Watch for whether companies add training and clearer rules for AI use, since the study suggests junior workers may be most at risk of losing bargaining power.

Source: Financial Times