Current data show AI has near-zero net employment impact
The recurring finding across recent research: no statistically significant relationship between AI adoption and unemployment or employment growth. Multiple independent studies, across different methodologies and data sources, all land in roughly the same place.
The Census Bureau found that only ~5% of AI-using firms reported any headcount impact at all — and those were split almost evenly between increases and decreases. The Atlanta Fed found that more than 90% of firms, on average, estimated no employment impact over the last three years. Yale's Budget Lab: "the picture of AI's impact on the labor market that emerges from our data is one that largely reflects stability, not major disruption." NBER: "AI adoption has not yet led to meaningful changes in total employment."
There is one genuine exception: entry-level roles with high AI exposure are harder to find. But researchers also found increases in entry-level roles where AI is augmentative, and unemployment for young SWEs has reverted toward trend — suggesting displaced workers found work elsewhere. That's reallocation, not apocalypse.
On the margin, "AI augmented" industries show stronger hiring and weaker unemployment than "AI substitution" industries. On earnings calls, AI-as-augmentation out-mentions AI-as-substitution by roughly 8:1.
AI Capex alone — data centers, electrification infrastructure — has likely made AI a net jobmaker already, before accounting for any downstream productivity effects. Construction and skilled-trades demand from AI infrastructure is expected to persist through at least end of decade.
Source claim: Across multiple independent 2025–2026 studies, AI's aggregate employment impact is near zero, with substitution and augmentation effects roughly balanced at the economy-wide level.