Coalmine Canaries

In the 19th century, coal miners brought canaries into tunnels as an early warning system. If the bird stopped singing or worse, dropped dead it meant toxic gases were seeping in. The miners had mere minutes to escape. Today, a new study from Stanford’s Digital Economy Lab suggests we’re seeing the modern equivalent in the labor market. The canaries are now young workers in their early 20s. The toxic gas is Generative AI. The paper, "Canaries in the Coal Mine? Six Facts About the Recent Employment Effects of Artificial Intelligence," uses payroll data from millions of U.S. workers to reveal a stark trend: since AI’s widespread adoption, employment for workers aged 22–25 in AI-exposed occupations has plummeted by 13%. 

Meanwhile, their older colleagues in the same roles remain largely unscathed. This isn’t just another "AI is changing work" headline. It’s evidence that the disruption isn’t coming. It’s already here and it’s targeting the most vulnerable first. At first glance, the findings seem counterintuitive. Shouldn’t AI threaten older workers clinging to outdated skills. Not necessarily. The study’s authors highlight a brutal paradox: Entry-level jobs exist, in part, to give workers experience. But AI is now experienced enough to skip the learning curve. Consider the roles hit hardest: Customer service reps replaced by chatbots. Junior coders outpaced by GitHub Copilot. Content creators competing with AI-generated drafts. These jobs have long served as on-ramps to careers. They’re where workers prove themselves, build skills, and climb the ladder. But AI doesn’t need a ladder. It starts at the top. Meanwhile, older workers in the same fields retain their jobs not because they’re immune to AI, but because they’ve moved into supervisory or hybrid roles that require judgment, mentorship, or complex problem-solving. Skills AI hasn’t mastered. Yet. 

The study delivers another gut punch: the adjustment isn’t happening through lower wages. Firms aren’t cutting pay to keep young workers employed. They’re cutting the workers entirely. This flies in the face of the optimistic narrative that AI will simply "augment" human labor, making us all more productive. For a 22-year-old whose first job just vanished, "augmentation" feels like a cold comfort. It also raises a terrifying question: If AI erases the bottom rung of the career ladder, how do young workers ever climb. History shows that labor market shocks don’t stay contained. 

Here’s what could come next: The Experience Gap Widens Without entry-level jobs, young workers can’t gain the experience needed for mid-career roles. This could create a "lost generation" of workers perpetually locked out of advancement while those who did secure early-career jobs before AI’s rise hoard the remaining opportunities. Sectors reliant on entry-level labor (e.g., tech support, digital marketing, paralegal work) may face talent pipelines drying up. If no one’s hired at 22, who’s ready to lead at 32. No one. 

For the Wrong Reasons Displaced young workers may flood into precarious gig work (Uber, TaskRabbit, Fiverr), where AI already dictates wages and opportunities. This would result in a a two-tiered labor market: a shrinking class of stable, AI-augmented professionals and a growing underclass of algorithm-managed gig workers. Entry-level jobs have long been a path to upward mobility. If AI closes that path, we risk entrenching class divides. Your parents’ wealth and whether they could afford to let you intern for free might become the only way to break into a career. It’s easy to blame "the algorithm" or "the march of progress." But the choices behind this disruption are human-made :Corporations deploy AI to cut costs, often without considering the long-term consequences for their talent pipelines. Policymakers have been slow to regulate AI’s labor impacts, treating it as a future problem rather than a current crisis. Educational institutions still churn out graduates trained for jobs that may not exist by the time they don caps and gowns. The Stanford study’s authors stop short of prescribing solutions, but the implications are clear: If we don’t act, we’re not just risking a few bad years for young workers. We’re risking a permanent underclass.