AI hiring – A new study of 21,559 U.S. firms finds that companies investing heavily in AI added workers over two years, while low-intensity adopters saw little headcount movement—countering fears that adoption automatically leads to layoffs. Yet researchers warn the pictu
For many Americans, the dread about artificial intelligence isn’t abstract—it comes with timing. On the days when hiring pauses, a familiar question follows: did AI take a seat at the table, and did it kick someone else out?
New research is pushing back on a central version of that fear. Firms that spent more on AI tended to grow their headcounts over a two-year period. while firms that spent little saw little change. The findings challenge the idea that “big adoption” inevitably means job losses. even as workers and employers alike remain unsure what happens next.
The data comes from researchers at Ramp and the workforce intelligence company Revelio Labs, who compared AI spending and employee data across 21,559 U.S. firms. The study split companies into “low-intensity” and “high-intensity” adopters, based on how much they spent on AI per employee.
Over the two years tracked, low-intensity adopters—firms that averaged $2.78 per employee on AI—saw little change in their head counts. High-intensity adopters—those averaging $33.67 per employee—grew their head counts by 10.2%.
The shift shows up not just in overall hiring, but in entry-level roles. Over the same period, entry-level head count at high-intensity firms rose 12%, directly contradicting predictions that young or inexperienced workers would be the first to lose out as AI rolls into workplaces.
The results arrive as anxiety is already widespread. A Reuters/Ipsos poll released in June found that 53% of Americans worry AI could cost them or someone in their household a job.
Researchers behind the study said their approach matters because much of the prior research focused on AI-exposed industries without comparing AI spending directly against staffing levels at the same firms. Ramp Economics Lab’s lead economist Ara Kharazian said the study’s design helps narrow that gap.
Still, the sample comes with limits. The research is not nationally representative, because it focuses on tech-forward companies and white-collar workers. It also does not include roles that may be affected in other ways—such as drivers potentially displaced by self-driving cars. or jobs lost through manufacturing automation. Kharazian said those missing categories mean the findings cannot be treated as a complete map of all jobs and all industries.
Even for the firms that were studied, there’s a timeline question. The researchers have not ruled out that head counts may change after the two-year window they tracked. They plan to update their results regularly as more data becomes available.
That uncertainty lands differently for job seekers. Kharazian said the unknown long-run impact puts entry-level workers in a bind today.
“If you’re an entry-level employee in the job market. or you’re just graduating college. it’s really hard to know what should I do. ” he said. “Join the firm that uses AI, and then worry about getting laid off because that’s the narrative out there?. Or do I join the firm that doesn’t use it at all. and then risk not learning this new technology that everyone’s telling me to use?”.
After crunching the numbers, Kharazian said he would steer an entry-level job seeker toward the company “going all in on AI,” because, in his view, the evidence they have points to jobs changing rather than disappearing.
ZipRecruiter economist Nicole Bachaud framed the shift in skills rather than simply headcount. “The one thing we know for certain is that jobs are changing,” she said. “A clerical or administrative clerk maybe needs to have more technical literacy today than they did five years ago. but a software developer maybe needs to have more interpersonal skills than they did five years ago.”.
Her comments reflect another tension running through the broader debate: Americans hear stories about layoffs and fear the technology is a shortcut to fewer jobs. But one October Yale Budget Lab study found the broader labor market had not experienced “a discernible disruption” since ChatGPT’s release. adding that widespread technological impacts tend to show up over decades. not months or years.
That broader market framing is one reason economists often sound more optimistic than the public. In January, a World Economic Forum report estimated that new technology, along with other economic and demographic trends, will create 170 million roles and displace 92 million by 2030.
Kory Kantenga. head of economics for the Americas at LinkedIn. said it makes sense that companies increasing AI investment also increase hiring. He offered three reasons: companies with more money to spend on new technology likely also have more to spend on labor; rapidly growing AI startup ecosystems naturally add roles in areas like human resources. marketing. and sales; and when AI supports new growth at existing companies. those businesses may need additional workers to pursue it.
Kantenga gave an example of an AI workflow used to research and identify potential clients. Once the AI agent produces a list of prospects, he said, “you need someone working in business development who can call them and actually make a deal.”
Even so, he urged people not to dismiss concerns about reskilling and development given how quickly AI is changing day-to-day work.
“That said, do I think that there’s a prospect that we’re staring down the barrel of 50% unemployment in two years?” he said. “No. I do not.”
For businesses, the other unanswered question is whether AI is delivering the productivity payoff executives promised. Bachaud said many companies have not yet measured what AI is actually changing.
“A lot of these businesses have yet to actually be able to measure what the impact is and how AI is actually leading to this massive productivity gain that people are expecting. ” she said. “We have this excitement. We have the capabilities. What is it actually doing. and how are we measuring that to really be able to determine: Is all of this worth it?”.
Kharazian’s findings suggest those struggles may not be evenly distributed. He said business owners shouldn’t be surprised if they aren’t seeing results right away.
“The vast majority of firms that use AI, according to our work, don’t see the results today,” he said. “What we found is that you only see the results when you use AI with some high intensity. There’s probably some threshold requirement that you have to use AI a certain amount. and you have to use AI for a sustained period before you start to see the results.”.
He added that successful deployment likely requires organizational change and employee-level learning so workers can use AI effectively.
But he also pointed to evidence that can help firms interpret what’s happening. “But if you’ve tried it and you haven’t found those gains and you’re wondering what you’re doing wrong, I’d say that is quantitatively backed,” Kharazian said. “But there is still growth to be found.”
AI hiring job prospects Ramp Economics Lab Revelio Labs workforce intelligence headcount growth entry-level hiring AI spending per employee automation fears job market 2026 Reuters Ipsos poll
So AI hiring doesn’t lay people off… but it will later, right?
I don’t buy it. Companies “add workers” now but then they’ll just fire the ones they don’t need once the AI learns everyone’s jobs. Happens every time.
They say the study is looking at AI spend intensity, but what even counts as “high-intensity”? Like if a company bought software for HR does that mean they’re hiring more cuz of AI, or cuz business is good?
This reads like they’re trying to calm people down, which is whatever. My cousin worked at a place that “used AI for scheduling” and suddenly the team got smaller even though they kept saying it was fine. Data might show headcount, but hours and pay can still get cut. Also, the headline says “until intensity changes” like that’s just a polite way to say layoffs when they feel like it.