AI's Effect on Work Shows Up in Wages as Young Workers Face 19% Employment Drop in Exposed Roles

AI's Effect on Work Emerges Through Wages, Not Mass Layoffs

AI's effect on work is manifesting in unexpected ways across the labor market. Despite predictions from Anthropic CEO Dario Amodei in May 2025 that AI could eliminate half of entry-level white-collar jobs within five years and push unemployment above 10%, the unemployment rate has actually fallen to 4.1% in August1

1

. However, new research reveals AI adoption by businesses is registering through slower wage growth and reduced job openings rather than widespread terminations.

Princeton researcher Sania Edlich and Torsten Slok, chief economist at Apollo Global Management, analyzed data across 321 occupations and found that workers in AI-exposed occupations experienced slower wage growth over the past three years while employment levels remained unchanged. Companies are capturing productivity gains by suppressing wages rather than cutting jobs. "If you are a journalist or an economist, there is just more pressure on wages because now the work that we do can be replaced, at least in some ways, by AI," Slok explained1

1

.

Young Workers Bear the Brunt of AI Transformation

The impact on employment for workers aged 22-25 has been particularly severe. Stanford's Digital Economy Lab released a working paper showing that young workers in AI-exposed occupations now face employment levels 19% below where they would be had they kept pace with their less-exposed peers2

2

. Economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen analyzed high-frequency ADP payroll data covering millions of U.S. workers through June and found no comparable gap for experienced workers in the same occupations.

Gen Z workers are scrambling to adapt. Bank of America Institute reported on September 9 that Gen Z's rate of switching jobs has overtaken every other generation for the first time since 2021, even as broader hiring slows. When they do switch, Gen Z receives the largest pay bump of any generation, suggesting they're being pushed out of AI-exposed roles and moving laterally into whatever positions remain available2

2

.

Wage Growth Collapses in AI-Exposed Occupations

The research by Edlich and Slok found that AI-exposed occupations have experienced a 6.7% decline in real wage growth since 2023, with the harshest effects concentrated among lower earners1

1

. Service workers in administrative or support roles have seen a 24% decline in real wages, while the bottom quarter of workers experienced an 11% decline. Top earners have barely been affected, revealing a stark divide in how AI impacts different income levels.

Melissa Krut, senior vice president at Sogolytics, a Virginia software company serving over 10,000 clients, described how productivity has increased and client numbers have grown over the past three years, yet proportionate hiring hasn't followed. AI models now handle more than half the baseline tasks for some jobs, leading to "higher expectations" for workers seeking raises and new hires. "If your job could be done by AI, yeah, you better have something else on top," Krut said1

1

.

Two Competing Futures: AI Substitutes vs. AI Complements Labor

The labor market is splitting into two distinct patterns. Gad Levanon, chief economist at the Burning Glass Institute, conducted an experiment ranking every industry's quits rate against its own 25-year history. In finance, insurance, information, and professional and business services (FIIPB), the quits rate has collapsed to the 13th percentile of its 25-year range, sitting at 1.8% compared to 2.5% in 2019—a 28% drop and the lowest reading since 20132

2

.

"Job hugging is real, but it's mostly happening in one part of the economy," Levanon wrote. FIIPB employment peaked in early 2023 and has been falling since. Bureau of Labor Statistics data from September 1 shows professional services hires fell by 188,000 in July alone, even as job openings ticked up nationally. Meanwhile, the rest of the private economy sits at the 44th percentile of its historical quits rate, close to normal2

2

.

Levanon attributes this to "a decline in the labor intensity of white-collar work." FIIPB output kept growing, but the labor needed to produce it didn't. Both Levanon and the Stanford team found the decline concentrated in occupations where AI substitutes for labor, while occupations where AI complements workers show flat or rising employment, especially for experienced workers2

2

.

The Tacit Knowledge Advantage

Economist Tyler Cowen has been drawing distinctions between raw "intelligence" that can be automated and tacit, contextual expertise he calls "Polanyi knowledge," after polymath Michael Polanyi. Stanford's payroll data confirms that codifiable routine jobs are disappearing specifically for workers with the least experience—those who haven't yet accumulated the human judgment and tacit knowledge that protects more experienced workers2

2

.

Loujaina Abdelwahed, head of economic research at Revelio Labs, warns: "We could be living in a world where unemployment is low while job quality, composition and wages deteriorate"1

1

. The mass retirement of baby boomers isn't creating an upward swell in salaries or job openings for Gen-Xers or millennials either, as companies opt to capture AI efficiency gains rather than expand their workforce proportionately.

Liminal Capital estimated in August that about a third of jobs appear more exposed to substitution than augmentation by AI. In those occupations, new hires among workers ages 22 to 25 have fallen by a third since 2021, even as firings have barely budged1

1

. The pattern suggests employers are becoming more selective, raising the bar for entry while maintaining existing staff—a shift that fundamentally alters how young workers enter professional careers and what skills they need to succeed.