AI and Entry-Level Jobs: What the Early Evidence Shows
- A Stanford analysis found a 16% relative employment decline among 22- to 25-year-olds in the most AI-exposed occupations.
- The researchers controlled for firm-level shocks but did not claim that AI explains every labour-market change.
- Less-exposed occupations and more experienced workers did not show the same pattern.
- Economic conditions and interest rates may also contribute to weaker recruitment.
Stanford Digital Economy Lab’s study of early-career employment points to strain where many careers begin. The researchers used administrative payroll data to compare employment patterns by age and occupational exposure to AI.
The evidence is suggestive rather than conclusive. The Stanford analysis found a 16% relative employment decline among workers aged 22 to 25 in the most AI-exposed occupations since generative AI became widely used, even after controlling for firm-level shocks. Employment in less-exposed fields, and among more experienced workers in the same occupations, remained stable or continued to grow.
Those figures do not prove that AI alone caused the change. Interest rates and wider economic conditions may also have reduced recruitment, while a technology changing this quickly is difficult to isolate. The pattern still matters because junior roles are more than cheap labour: they are where people learn the routines, judgement and professional relationships needed for later responsibility.
Software development and customer service are among the occupations highlighted in the research. Completing a bounded task is different from being accountable for its consequences. A system may produce plausible code, a customer response or a financial draft, yet carry no professional responsibility when the work is wrong.
The question is which organisations decide that human checks, training and accountability are worth paying for once routine tasks can be automated.
This matters immediately for education. If entry-level work becomes scarcer, courses and employers will need clearer ways to show what a learner can do independently, where AI assisted, and how the final work was checked. Honest disclosure is not a bureaucratic ornament; it makes skills visible when polished output can be generated cheaply.
The risk is not simply fewer jobs in a spreadsheet. It is a thinner route into occupations that depend on accumulated experience. The public may then discover too late that efficiency savings removed the people who learned to spot a costly mistake.