Amid the A.I. boom, warning signs are flashing for the labor market

New data and recent studies indicate that A.I. may be causing joblessness.

Is A.I. beginning to affect the job market? A couple of studies that were released in the last week say so.

A Stanford analysis, somewhat provocatively titled “Canaries in the Coal Mine?”, found a 13 percent decline in employment rates among early-career workers in sectors most vulnerable to A.I. And researchers at the St. Louis Fed found that higher rates of A.I. adoption by industry were directly correlated with an increased rate of unemployment for workers in those sectors.

Does this answer the ever-burning questions — and debates — about how A.I. will reshape the jobs market? Of course not. But these two studies seem to be early indications that the implementation of the technology is already increasing joblessness in some ways.

Both the St. Louis Fed and the Stanford reports noted that the technology sector is, unsurprisingly, one of the industries with the highest rate of A.I. adoption — and therefore employment impacts. Other reports and anecdotes seem to back this up. Tech giant Oracle disclosed another round of layoffs this week as it ramps up its spending and investment in A.I. products, on top of a previous round of cuts in August. And Salesforce CEO Marc Benioff told some podcasters that he had recently eliminated 4,000 roles due to A.I.

Some of the most important measurements of the labor market appear to show a slowdown in hiring too. The jobs reports for the last three months have totaled 106,000 together — a paltry average of about 35,000 new jobs added a month. That brings the average for the year down to about 85,000 jobs a month, about half of last year’s average of 168,000.

Meanwhile, more federal labor data released this week, the JOLTS survey, showed that the U.S. now has more unemployed people than openings for the first time since April 2021 — right around the time vaccines were getting rolled out en masse. That kicked off a period where there was more work to go around than workers, causing business headaches around labor shortages but also helping lift wages in the process. My former WaPo colleague Heather Long, who is now the chief economist at the Navy Federal Credit Union, wrote that this week’s JOLTS data was “yet another crack in the labor market that illustrates how much harder it is to get a new job right now than what we've seen in a long time.” CNN’s slightly overcooked take? "Confirmed: America is in a serious jobs slump.”

The slowdown, should the trend continue, is a result of a multitude of factors. But the St. Louis Fed report appears to be one of the strongest indicators that A.I. is at least a part of that equation.

“Our results suggest we may be witnessing the early stages of AI-driven job displacement,” researchers Serdar Ozkan and Nicholas Sullivan noted in the report. “Unlike previous technological revolutions that primarily affected manufacturing or routine clerical work, generative AI can target cognitive tasks performed by knowledge workers—traditionally among the most secure employment categories.”

A previous paper from the St. Louis Fed noted that some 23 percent of workers are using generative A.I. at least once per week.

Meanwhile, the Stanford study drew an important distinction from their data: between industries that are using A.I. to augment work, which has not resulted in job loss, and those that have been using A.I. to automate work, which does.

They reached their conclusion by finding that young adults experienced the same employment growth as older workers in fields that had not been shaken up by A.I., across the board. But in fields where A.I. is being most readily deployed, the employment rate for adults aged 22-25 dropped by 6 percent from late 2022 to July 2025, while older workers saw their employment rate increase 6 to 9 percent.

Here’s what else we’re reading this week:

-A nice audio report from WBUR goes deep on looking at how A.I. is increasingly being used in the hiring process by companies, to screen applicants or even conduct interviews. A college student applying to jobs told the outlet that she was interviewed by an A.I. bot for the job. By the third question, it just started saying “vertical bar pilates,” over and over again, a maddening loop that she recorded, and played on the segment.

AI BOT: Vertical bar pilates. Vertical bar pilates. Vertical bar pilates. Vertical bar pilates. Vertical bar pilates. Vertical bar pilates. Vertical bar pilates.

She never head back about the job.

-If anyone wants to get into the weeds of labor policy, this interview with Teamsters president Sean O’Brien, who drew sharp criticism for his decision to speak at the RNC and sort of tacitly support Trump, blew my mind. I guess he felt like he was on friendly ground with the Free Press, they really got him going. We can unpack that more next week.

-If this Reddit post is legit, and it looks like it is, the San Francisco Office of Labor Standards Enforcement may be looking into labor practices at Scale AI, the annotation and A.I. training startup that Meta invested heavily in earlier this summer. We wrote about them in July. On Reddit this week someone who said they were an investigator with the department posted a note looking to talk to people in San Francisco who had done annotation work for the company. Interesting…

-Former CNN media reporter Oliver Darcy talks with Peter Kafka about why the mainstream media doesn’t get it — and why he wanted to build his own outlet.

-And Kafka wrote a piece looking at the mere slap on the wrist Google received after being found, after a ten-week trial in 2023, to be engaging in monopolistic behavior regarding the dominance of its Chrome browser. The case was perhaps the biggest victory for anti-trust advocates and government lawyers, who argued that the company should be forced to sell Chrome. Kafka said that the whimper-like ending of the antitrust effort signified that the era of “the Big Tech backlash is formally over.”

Google investors have spent years mostly ignoring the idea that the company could face severe punishment from US regulators. This week confirmed that they were correct. So maybe it's time to formally acknowledge that the larger push in the US to break up, or simply rein in, Big Tech is basically over…We saw what happens when the US government actually wins a court case against a Big Tech company — not that much at all.

See you all next week.