A 70% factory AI pay gap is not a raise for workers

What the note measured

The 30 September FEDS Note uses Lightcast online job ads for U.S. manufacturing. Broad AI-skill requirements rose to about 11% of manufacturing postings in late 2025, compared with 8% across sectors. Explicit generative-AI skills remained under 1% of manufacturing ads through July 2026; machine-learning requirements account for much of the broad increase. Production occupations showed the same direction at lower levels.

Across manufacturing, ads requiring AI skills advertised about 70% higher average wages than all manufacturing ads over the sample. Among production occupations, the average gap was about 30% from 2023. These compare different posted positions. They do not show that a worker received either increase.

What the pay gap cannot answer

The wage comparison does not hold occupation, employer or location fixed. AI-skill ads are a small, changing group. Wages appeared in roughly 10% of postings before the pandemic and about 50% recently, so the observed group also changed. Online ads are recruiting activity, not unique openings or completed hires. The note does not link ads to workers' actual pay, employment, hours or layoffs, and it does not identify AI as the cause of the gap.

Why it matters

A low task-exposure label for production work does not mean manufacturers never ask for AI skills. But a skills request is not proof that generative AI is replacing production staff. The next check is whether comparable applicants get these jobs at higher accepted pay, and whether incumbent pay or staffing changes after employers adopt AI. Until then, this is a change in advertised demand and an unadjusted price difference, not a personal career forecast.