According to a St. Louis Fed study released last week the U.S. “may be witnessing the early stages of AI-driven job displacement,” with a weighting toward the sectors which adopted the emerging technology most heavily.

The research. released on August 26, sought to establish whether AI is contributing to rising unemployment. This comes after an unwelcome surprise in the labor market early last month when the Bureau of Labor Statistics hugely revised down its data: May’s tally was cut from 144,000 to 19,000, and June’s total was slashed from 147,000 to just 14,000, resulting in a combined loss of of 258,000.

The weaker picture of the economy prompted a raft of questions: Is hiring slowing because of fears over Trump’s tariff plan? Is the employment market slowing because of uncertainty more widely? Or is there a factor which is fundamentally reshaping the labor market?

We have also heard the many, many warnings about jobs displaced due to AI. Is there a possibility that this is driving the underlying shake up?

“According to the nationally representative Real-Time Population Survey (RPS), 23% of employed workers used generative AI for work at least once per week as of late 2024—a remarkable adoption rate for such a nascent technology,” wrote the St. Louis Fed’s (FRED) Serdar Ozkan and Nicholas Sullivan. “Despite this widespread integration, we still know surprisingly little about AI’s employment effects because of the newness of the technology.”

What the FRED can chart is the percentage point change in unemployment between 2022 and 2025 in certain industries, and its correlation to AI exposure in each of the sectors.

The research showed a correlation coefficient of 0.57, meaning generally the occupations that embraced generative AI most intensively showed the largest unemployment increases. Those sectors included, at the extreme end, computers and math. In these professions, AI adoption was at a little under 80% while the unemployment change increased by 1.2% over the past three years.

Of course, if you’ve checked in on tech employment over the past three years AI hasn’t been the only story in town. Big Tech especially was criticized for overhiring during the pandemic, prompting a wave of layoffs in the years following.

Former PayPal boss Keith Rabois, for example, said in 2023 the axing of many roles was overdue: “All these people were extraneous, this has been true for a long time, the vanity metric of hiring employees was this false god in some ways … There’s nothing for these people to do—it’s all fake work. Now that’s being exposed, what do these people actually do, they go to meetings.”

Likewise tech specialists—particularly those in the AI field—told Fortune they were being paid six-figure sums to be “penned” in by certain companies in order to stop rivals hiring top talent. Yet by hiring these individuals with no real job for them to do, the employees often ended up doing a 10-minute task a day before using their working hours as free time.

Big Tech didn’t try and hide the correction either. Mark Zuckerberg launched his “year of efficiency” in 2023 which shrunk headcount by 22% after years of double-digit growth, with Alphabet’s Sundar Pichai adding in 2024 that Google would be “removing layers to simplify execution and drive velocity.”

Safer harbors

At the opposite end of the spectrum, industries with lower AI adoption rates are seeing relatively unchanged employment levels. The personal services industry, for example, had the lowest adoption rate of the sectors surveyed and had an unchanged employment rate.

Likewise, the legal and social services sectors had an adoption rate of around 18% and negative unemployment rates over the past three years.

There’s also evidence to suggest that AI can be more disruptive to careers depending how recently a person has joined the labor market. For example, a landmark study led by Stanford Professor Erik Brynjolfsson last month found entry-level workers in the occupations most exposed to AI are already experiencing a 13% relative decline in employment.

Deutsche Bank, referencing Brynjolfsson’s study, noted to clients this morning: “It’s one of the first high-profile reports to identify the effects of AI potentially showing up in labour market data. It finds that since the launch of ChatGPT in November 2022, there has been a 6% decline in employment for workers aged 22 to 25 in the occupations that can most be augmented by AI—such as software engineering and customer services—even after controlling for firm- specific shocks.

“By contrast, there has been a 6% to 9% increase in employment for more experienced workers in the same professions, the study found, citing payroll data.”

Goldman Sachs also noted a change in hiring due to artificial intelligence in a research note Monday—but not because of displacement. The Goldman Sachs Analyst Index for August found 58% of surveyed analysts reported the companies they cover are hiring at about the same pace as at the beginning of 2025—but concentrated in AI-related positions. Conversely, companies were pausing or axing headcount for non-AI-related roles.