How will AI reshape our economies? What will the future hold for workers, and which occupations will continue to grow in the coming decade? These are the questions that everybody is asking right now. But none of us has a definitive answer. Researchers around the world are currently investigating AI’s impact on the labor market. The verdict: It’s complicated! For now, we don’t even know for sure whether AI is going to create or displace jobs in aggregate in the years to come. Here is what we do know, though: The technology is changing several pockets of the labor market — and often in unexpected ways. What we see in the data is that AI has started to create a three-tier labor market, with each of the segments performing quite differently. Let me explain.
The resurgence of blue-collar jobs
The biggest impact AI is having on the economy so far comes through the data center construction boom. Economists estimate that the amount of capital flowing into the AI economy is now surpassing previous technological cycles, including the railroad buildout during the 19th century. And the labor market is starting to notice. Job postings data shows that hiring demand for data center construction roles — HVAC technicians, electricians, and skilled trades in general — has gone through the roof across the country. This is particularly true in the remote areas where some of the construction is happening. Companies are fiercely competing for workers, recruitment costs are shooting up, and so is compensation. Higher salaries are required to lure blue-collar workers to the locations where the work needs to be done. A byproduct of this economic trend has been a decline in wage inequality as compensation in many of the skilled trades has outrun that of the average white-collar worker for several years in a row now.
Skilled AI workers: supply is not keeping up with demand
The second pocket of the labor market, where demand is soaring even faster, is skilled AI workers. By that, I do not just mean prompt engineers but rather white-collar professionals across various occupations who have acquired AI knowledge: It could be workers in sales or marketing, economists, consultants, data scientists — you get the idea. Job postings across advanced economies show that the share of job ads explicitly mentioning AI skills has more than tripled in recent years. In the U.S., the share went from about 2% in early 2023 to almost 7% more recently based on data from Indeed. With that increase in demand, compensation has also risen rather handsomely. Nowhere is this more visible than in job ads from big AI companies such as Anthropic, where practically every role pays remarkably well. They are giving Wall Street a run for its money! With AI being increasingly deployed across industries, the need for AI skills in various occupations will keep rising. As it takes time for the workforce to adapt and reskill, demand for workers with AI knowledge will continue to outrun supply. And that means labor market prospects for those workers should remain strong.
Routine white-collar workers are struggling, especially the young
The one big segment that is struggling is routine white-collar workers. While there haven’t been massive job losses, hiring has been depressed in recent years, and workers are stuck. This is affecting quite a few roles in sales, customer service, administrative support, human resources, content creation — the list goes on. What we’re seeing is that senior workers are mostly still doing OK, while graduates and junior workers are suffering. The hiring drought is affecting the young much more than the old. There are several explanations as to why that might be. Let me highlight two of them. First, AI is now doing the routine work that many junior workers would have previously done on the job. Meanwhile, it doesn’t touch more senior tasks: team leadership, oversight, communication with stakeholders, be they managers or clients, etc. Second, the higher prevalence of remote and hybrid work has made it less economical to hire at the entry level. Management and upskilling become more costly, while AI can potentially take on some of the work previously done by younger staff. As neither AI nor hybrid work is going away any time soon, this weakness could persist for a while. Both governments and companies will have to address this problem or eventually face a broader societal backlash.
Where does this leave us?
AI is certainly going to displace many workers, even as it creates a lot of new jobs. While the aggregate effect is uncertain, I am positive that the job Armageddon that some have predicted can be avoided. Nevertheless, the transition might be painful for some. We are already seeing how AI has started to change today’s labor market. No big net job losses, but certain groups are more severely affected than others. This includes young workers who have recently struggled to get a foot in the door, since AI seems to be complementary to skills associated with seniority and experience. Governments and the private sector alike need to think about how to support those who are negatively affected by AI deployment. This calls for better labor market institutions (reskilling and re-education programs) but also better recruitment practices: From a workforce planning point of view, it might not be the best idea to stop hiring young people when many older workers are about to retire in the decade to come. Just a thought!
