News

What Labor Data Actually Says About AI and Hiring

Most headlines about AI and jobs exist on one of two extremes: mass displacement is imminent, or nothing has really changed. A fresh look at the payroll data suggests both narratives are wrong — and knowing where the actual fault lines are matters enormously for how you position yourself in today's job market.

A working paper from the Stanford Digital Economy Lab offers one of the most granular readings of AI's employment effects to date. The paper, authored by Erik Brynjolfsson — the lab's director — along with Bharat Chandar and Ruyu Chen, drew on high-frequency payroll records from ADP spanning late 2022 through mid-2025. The headline finding: early-career workers in AI-exposed occupations took a measurable hit, while the broader labor market largely held steady.

Specifically, workers ages 22 to 25 in occupations most exposed to AI — including software engineering, marketing, and customer service — experienced a 16% relative decline in employment compared to workers in less-exposed roles. Within individual firms, entry-level hiring in those same AI-exposed jobs fell 13% relative to comparable less-exposed positions. As Time reported in its coverage of the study, Chandar noted the findings are "consistent with the hypothesis that AI is having this effect, especially for entry-level workers." The researchers stopped short of declaring definitive causation, a distinction worth noting.

The generational split in the data is striking. Workers aged 30 and older in the highest AI-exposure categories saw employment grow 6 to 12% during the same period. The divergence suggests AI is displacing junior workers in roles heavy on pattern recognition and routine tasks — precisely what large language models handle most fluently — while more experienced professionals, who bring contextual judgment and relationship capital, have so far been largely shielded.

Equally important is what the data does not show. According to the Stanford Digital Economy Lab's summary of the research, the overall impact of AI on aggregate employment is described as "likely small right now," with studies drawing on Current Population Survey data showing "at most minimal changes" in hiring across AI-exposed jobs at the macro level. The disruption, in other words, is concentrated rather than economy-wide — and the researchers themselves acknowledge "enormous uncertainty about how AI will shape labor markets" going forward.

What this means for job seekers

The data creates a clear, if uncomfortable, map. If you are early in your career and your target roles sit in software development, digital marketing, or customer service, the competition for entry points has genuinely tightened. That is not pessimism — it is the signal your job search strategy needs to factor in.

The more actionable finding is the augmentation vs. automation divide. Workers whose roles position AI as a tool — using it to accelerate output, manage complexity, or extend their expertise — are seeing stronger employment outcomes than those whose roles AI simply replaces. That means the strategic play is not to avoid AI-adjacent work but to build skills that sit above what the model can do alone: client relationships, domain judgment, cross-functional coordination. The researchers note that employment falls among workers using AI to automate tasks but grows for those using it to learn new skills — which makes continued upskilling less optional and more load-bearing for anyone navigating a career now.

The honest read: AI's labor-market impact has arrived unevenly and earlier than aggregate statistics suggest, concentrated most sharply at the career entry point. That asymmetry is what makes understanding the specifics — rather than the headlines — so valuable right now.