Is AI Really Taking Our Jobs? The Evidence Behind Jensen Huang’s ‘Lazy’ Critique

But does the data support Huang's confident dismissal, or is he selling a self-serving narrative while AI's effects quietly take root? Here’s what the latest research tells us.

The Case for 'Aggregate Calm'

Several major studies support Huang's claim that a broad "AI jobs apocalypse" has not arrived. Researchers at the Yale Budget Lab analyzed the labor market in the 33 months following ChatGPT's release and found "no discernible relationship" between measures of AI exposure and changes in employment or unemployment, concluding that the broader market has not experienced a meaningful disruption . A recent Brookings Institution review corroborated this, finding the current evidence on labor-market impact to be "inconclusive" and warning that specific claims of harm are "premature" .

Other key findings reinforce this picture of aggregate stability:

• ITIF analyzed U.S. data and found AI created roughly 119,900 direct jobs in 2024—including machine-learning engineers and data scientists—while explicitly AI-related layoffs stood at just 12,700, representing only 0.1% of all layoffs that year .
• Goldman Sachs Research projects that unemployment will increase by only half a percentage point during the AI transition period and identifies three distinct categories of new jobs created by the technology: infrastructure buildout, demand for AI-proficient workers, and entirely novel roles .
• The World Economic Forum’s Future of Jobs Report 2025 projects that while 92 million jobs may be displaced by 2030, 170 million new roles will emerge—a net gain of 78 million positions .
• A new MIT Sloan study tracking AI adoption from 2010–2023 found the impact is typically on tasks within jobs rather than entire occupations, and employment can even grow when AI only automates a fraction of a role .

The Evidence for 'Localized Pain'

While the overall picture looks calm, a growing body of more granular research reveals that AI is already leaving a mark on specific groups and occupations—with early-career workers hit hardest.

A landmark working paper from the Stanford Digital Economy Lab, using high-frequency ADP payroll data, found that early-career workers (ages 22–25) in AI-exposed occupations experienced a 16% relative employment decline. Employment for experienced workers in the same fields held steady, suggesting that companies are cutting junior hires while retaining senior staff .

Other studies point to a widening crack in the foundation:

• The Federal Reserve Bank of St. Louis identified a "striking correlation" between AI exposure in occupations and local unemployment increases since 2022, with a correlation coefficient of 0.47 .
• Anthropic introduced an "observed exposure" metric that combines theoretical AI capability with actual real-world usage data. Their research finds that while AI is "far from reaching its theoretical capability," occupations with higher observed exposure are projected by the BLS to grow less through 2034 .
• Outplacement firm Challenger, Gray & Christmas reported that U.S. employers explicitly cited AI as the reason for roughly 55,000 job cuts in 2025, representing 4.7% of the 1.17 million total announced layoffs that year .
• The International Monetary Fund estimates roughly 40% of jobs globally—and closer to 60% in advanced economies—face meaningful exposure to AI capabilities .

Huang's Argument: Insightful or Self-Serving?

Critics might note the obvious: Huang runs the company that makes the chips powering the AI revolution. A narrative that AI kills jobs is bad for business. His framing—“if the world runs out of ideas”—flips the question from one of technological capability to one of human imagination, a convenient rhetorical move .

Yet even the most rigorous researchers largely validate his core temporal claim. At the aggregate level, the disruption skeptics feared simply is not visible yet. The data points not to a sudden jobs crisis but to a gradual, uneven shift affecting the youngest and most vulnerable workers first. The evidence is best summarized as early, localized pain amid aggregate calm, with most economists agreeing that the true impact of AI on the labor market will only become clear over the course of years, not months .