A new Stanford working paper finds that generative AI is reshaping the labor market by hitting early-career workers hardest. Using payroll data from ADP covering 25 million employees, researchers report a 13% employment decline for entry-level staff in AI-exposed fields since 2022. The steepest drops—about 20%—occurred among junior software engineers and customer-service workers, even as headcount for older peers in the same roles increased. Younger workers appear more substitutable for large language models trained on text-based knowledge, while experienced employees benefit from tacit and communication skills that are harder to automate. Occupations showing similar pressures include accounting, secretarial work, programming, and sales. By contrast, roles where AI augments rather than replaces labor—such as nursing—haven’t seen comparable entry-level declines. The findings underscore a shift toward AI-enabled productivity that rewards workers who can harness the tools—and penalizes those whose tasks AI can already do.
Related articles:
Generative AI at Work
AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity.
