What is happening to white-collar entry-level jobs? Ask around and you’ll hear a familiar story: AI will automate at least half of them within a few years, and their death is only a matter of time. I’ve spent most of the past year in the data, and the picture is more complicated. Entry-level jobs are changing quickly, but reports of their death are greatly exaggerated.
What Do We Know?
The slowdown started before AI. In spring 2022, the Federal Reserve began aggressively raising interest rates to fight inflation. Since then, labor market churn (job openings, hires, quits, and layoffs) has fallen substantially. The result is a low-hire, low-fire economy: employers hire sparingly and rarely cut, so employment remains historically high and unemployment historically low.
Entry-level hiring has followed the same path, only more steeply, because entry-level workers are usually the first to feel a shift in employer behavior. They are the last hired and the first fired. Their work tends to be routine, well-specified, and easy to substitute for. And they take time to become productive, so they are less an immediate payoff than a long-term investment, which makes them easy to postpone when budgets tighten.
When entry-level job postings began to fall, many people blamed AI. But that story had a problem from the start: the decline began well before ChatGPT’s release in November 2022, and long before most employers had built AI into their workflows. More likely contributors are a correction after pandemic over-hiring, higher interest rates, and a more uncertain policy environment. The story has a second problem: what employers themselves are saying. A March survey of nearly 1,500 U.S. executives and senior talent leaders found that, among organizations using or exploring AI, leaders were 2.7 times as likely to expect AI to raise their entry-level hiring in 2026 as to lower it. The firms with AI built into daily operations were the most upbeat.
Is AI Different This Time?
So far, AI looks like a normal technology. It automates specific tasks and frees workers for higher-value ones. Since the Luddites smashed mechanized looms more than two centuries ago, every generation has had its “the machines are coming for us” moment, from the telephone and the assembly line to the ATM and the spreadsheet. Spreadsheets were supposed to finish off accountants, and the field has added more than a million jobs since. Nor is today’s pace of change unusual. Economist David Deming and his colleagues found that the most turbulent thirty years in American job history ran from 1940 to 1970, while the 1990s through the 2010s were among the calmest on record. They also find that change has picked up since the pandemic, so this is a reason for calm, not complacency.
Technology typically creates more jobs than it destroys. Existing jobs evolve, and entirely new industries appear. E-commerce, cloud computing, software-as-a-service, the gig economy, and cybersecurity barely existed a few decades ago, and they now employ millions. What industries will generative AI create? Time will tell.
How Is Entry-Level Work Changing?
Employers describe two effects at once. In the March survey, 41 percent of leaders said routine tasks are shrinking, and 42 percent said the analysis and judgment expected of new hires is growing. That is why “AI replaces entry-level work” covers only half the picture.
Think of the traditional first job as a bundle: some grunt work, some learning, and a first look at how an organization runs. AI is pulling the grunt work out of the bundle, and what’s left is harder. Employers increasingly expect it from day one. PwC’s analysis of 2.4 million job postings points the same way: entry-level roles with high AI exposure were seven times as likely to demand skills once reserved for senior staff. Those reshaped roles grew 35 percent between 2019 and 2025, while other entry-level roles shrank 10 percent.
A higher bar might suggest employers want a new set of skills, but they mostly want a stronger version of the old ones. In the March survey, critical thinking, communication, and collaboration were the most sought-after skills, while AI skills ranked dead last. Employers seem to be saying that new hires already have the AI skills down, or that anyone with strong foundations can pick them up as needed.
What has changed is how employers separate the job-ready from the rest. Resumes, interviews, and GPAs tell them less than they used to. Increasingly, they lean on relevant work experience, whether from an internship, a regular job, or project-based work, along with behavioral and situational assessments and referrals from trusted sources.
What Don’t We Know?
We are in the early innings, and nearly everything above comes from the era of copilots, in which AI assists a person. Agents, which act on their own, may change the picture, and we have little data on them. One question worries me the most: If employers prize relevant experience above every credential, who is responsible for providing it?
What Should HR Leaders Do?
Rather than automating entry-level work to cut headcount, rework job descriptions and postings so that the tasks and skills reflect what new hires can now do, which is more analysis and judgment, earlier. Screen for relevant experience over pedigree. And build practice in deliberately: routine work used to train people as a side effect, so simulations, structured mentoring, and rotations will have to do that job now. The employers I’d bet on are the ones who treat this as a design problem, not a cost-cutting opportunity.
The first rung isn’t disappearing, but it is being rebuilt higher up. Organizations that rebuild it on purpose, instead of waiting to see what happens, should come out ahead.
