I've spent my career in and around the HR field, and one of the things that experience teaches you is to pay attention to early signals—the data points that don't make the headlines but tell you something important is shifting beneath the surface. So, when researchers at Stanford's Digital Economy Lab published a study last fall tracking employment outcomes for workers ages 22 to 25 in AI-exposed occupations, I couldn't stop thinking about it.
The finding was specific and striking: a 16% relative employment decline for early-career workers in those roles, while experienced workers in the same occupations held steady or grew. The researchers called these workers "canaries in the coal mine"—the first to experience what others haven't yet.
Entry-level roles have always done two things at once. They produce output, yes, but more importantly, they produce people. The junior employee processing claims, writing first drafts, building initial analyses, handling tier-one customer calls—that person is learning something that cannot be taught in a classroom or replicated in a training module. They are building judgment. They are accumulating the kind of tacit knowledge that only comes from doing real work, making real mistakes, and being close enough to someone more experienced to absorb how they think.
Generative AI is now very good at the procedural, checkable layer of that work. And organizations are reasonably drawing the logical conclusion: if AI can do it faster and cheaper, why hire someone to learn on the job? That logic misses what is lost in the bargain. Strip out the entry point, and the process by which organizations grow experienced people is interrupted, and the apprenticeship engine breaks.
I've watched this dynamic play out in other contexts. When organizations eliminate the roles where people learn, they discover the cost years later, when there's no internal pipeline ready to step into senior positions. The talent was never developed because the conditions for developing it were removed.
To be fair, the causation here is contested. Yale's Budget Lab finds no clean economy-wide link between AI and unemployment. June 2026 research from the New York Fed attributes a significant portion of the rise in young-graduate unemployment to remote work rather than to AI directly. According to that study, training remote workers is harder than training workers based in a physical office. The Stanford researchers themselves acknowledge their findings may reflect factors beyond generative AI alone.
Whether the mechanism is AI substitution, the difficulty of remote onboarding, or some combination, the direction is the same: early-career workers in high-exposure occupations are losing ground. Something is happening at the entry point of the workforce, and organizations that treat it as an abstraction rather than an organizational decision are going to feel the consequences.
Earlier this year, IBM's CHRO announced the company was tripling its entry-level hiring, explicitly including software developers, one of the occupations the Stanford study flagged. Her reasoning stayed with me: the companies that will be most successful three to five years from now are those that doubled down on entry-level hiring in this environment. AI can boost productivity, she said, but it cannot develop the next generation of technical leaders or innovators.
That's an organization choosing to protect its pipeline, even when the short-term math might argue otherwise. Not every company has IBM's resources to make that bet at scale. But every organization is making some version of it right now, whether they've thought about it in those terms or not.
The companies quietly reducing early-career headcount for efficiency are also placing a bet: that the pipeline can be rebuilt when they need it, that experienced talent will be available externally, that the apprenticeship engine can be restarted after being shut down.
I talk a lot about aligning organizational decisions with stated values. The decision on whether or not to invest in early-career talent reveals what an organization truly believes about how people grow.
About the Author Amy Schabacker Dufrane, Ed.D., SPHR, CAE, is CEO of HRCI, where she drives strategic conversations about building high-performing HR teams. An award-winning thought leader at the intersection of talent strategy and continuous learning, Dr. Dufrane is a celebrated keynote speaker on the human side of successful business strategy.
Related Learning & Resources
Certificate in Managing the Hybrid Workforce