The story begins in the second half of 2022. That is when Goldman analysts noticed something shifting in the data.
Fields with more exposure to AI started seeing slower growth in job postings. The pattern showed up most clearly in the U.S., Germany, and Australia. It was not a crash. It was a slowdown, a quieter kind of pressure.
Information and communication services, which includes software and media, has seen hiring growth decelerate in nearly every major advanced economy since 2022. Outside the U.S., staffing in that sector still sits at or above long-term norms. So the sector is not collapsing. It is just not growing the way it used to.
For the broader labor market, the numbers that are larger, just call. For every 10% increase in AI exposure, the yearly growth in staffing dips by just 0.1 percentage point in France, Canada, and the U.S. That is barely a blip.
Some industries are feeling this much more sharply than others.
Across developed economies, sectors such as customer support, software development, advisory services, and marketing have seen staffing levels well below historical norms. These are fields where AI tools can already do a meaningful chunk of the work.
The call center numbers are the most striking. Employment in call centers is 39% below trend in the U.S., 33% below trend in Canada, and 27% below trend in Germany.
Goldman's conclusion is that the workforce effects of automation are no longer theoretical. The tools are not hypothetical. They are deployed, and the hiring data reflects it.
The most interesting part of the research is about who feels this first.
The study examined employment patterns across a broad set of occupations, spanning more than eight hundred different roles. The data showed that early-career positions experienced the most significant negative effects from AI adoption.
Companies are not firing their experienced people to make room for AI. They are simply not hiring as many new people to backfill and expand. New employees are beginning to feel the effects of entry-level pressure.
Here is the bigger picture. By merging 11 national data, Goldman estimates that AI adoption in major developed economies sits between 15% and 20%, with France, the U.S., the Netherlands, and the U.K. at the forefront.
Italy, Japan, and New Zealand are along the lower end. Major emerging markets are at 10% to 15%.
The hiring effects are visible but confined to a fairly small set of industries and workers. That could stay true as long as the trend extends, or it could expand as adoption climbs.
For your portfolio, this is a reason to pay attention to which companies are actually using AI to cut costs and which ones are just are talking about it. The ones with real AI adoption are the ones showing up in these numbers.
For your career, the message is more personal. If you are early in your working life, AI exposure is not abstract. It is showing up in the hiring data.
The good news is that the overall effect is still small.
