Federal Reserve Governor Michael Barr said the United States should prepare for “serious short-term disruptions” in the labor market as artificial intelligence spreads, warning that the pace of deployment will determine whether the technology displaces workers at scale and widens inequality in favor of investors.
In remarks prepared for delivery at Yale University on Thursday, Barr did not forecast specific job losses but argued that preparing for social strains tied to AI is prudent. His comments follow midterm elections that swung partly on a backlash to the rapid AI buildout, and they precede a public Fed forum in San Francisco on Monday convened by a task force appointed by Fed Chairman Kevin Warsh to explore what the new technology may mean for the economy.
Barr framed the implications as deeply uncertain, with much riding on whether competition among AI companies continues to the point that the powerful new technology becomes “a cheap and ubiquitous commodity, with gains widely shared,” much like the internet became over time. In that scenario, he said, “AI could democratize access to skills and information needed for small businesses to thrive.”
The alternative, he warned, is increasing market concentration among those who own or have invested in a handful of winning companies. In that case, gains would be concentrated and increasingly flow toward owners of labor-displacing capital, continuing a downward slide in labor’s share of national income.
That share, a rough proxy for how much of the country’s output translates into wages for workers versus profits or equity income for business owners and investors, hit a record low of 52.8% in the second quarter of 2026. It had held roughly steady between 60% and 65% from World War Two until early this century.
Barr pointed to research suggesting that AI has characteristics that, in past examples of high-tech advances, reinforced concentration of market power. Even so, he noted there has been “relatively little displacement” of workers so far, and even evidence of AI boosting demand and adding jobs. But he said the issue remains unsettled, hinging in particular on whether small businesses can maintain access to the technology to expand.
“No one knows the degree to which the benefits will be shared,” Barr said. “We are already seeing signs that AI may be a labor substitute in some sectors, especially for younger, less-experienced workers ... If labor market changes happen quickly it will be hard for workers to adjust, and dislocations might be large ... The concentration of market power may help drive a widening of inequality that would strain the American Dream.”
The remarks sketch two divergent paths. Under one, robust competition keeps AI accessible and affordable, allowing small businesses to adopt tools that broaden opportunity and support job creation. Under the other, a small set of firms captures the gains, reinforcing their market position and channeling a larger share of national income toward capital owners rather than workers.
Barr’s focus on small business access underscores a key variable: whether the technology remains within reach for firms that lack the scale to develop their own systems. If access narrows, the benefits of AI could accrue disproportionately to larger players, further tilting the competitive landscape.
The labor share data he cited provides context for that concern. The measure’s decline from its mid-century range to a record low of 52.8% in the second quarter of 2026 illustrates how the split between wages and profits has shifted over time. Barr’s warning is that AI could accelerate that shift if deployment outpaces workers’ ability to adjust.
He also flagged the particular vulnerability of younger, less-experienced workers, who may face greater difficulty adapting if labor market changes arrive quickly. That concern aligns with his broader point that the speed of adoption matters as much as the direction.
The Fed’s involvement, including the task force appointed by Warsh and the upcoming forum in San Francisco, signals that policymakers are examining AI’s economic implications even as the evidence remains limited and mixed. Barr’s remarks stop short of predicting a specific outcome, instead emphasizing preparation for a range of possibilities.
For markets and sectors, the remarks point to potential effects across technology, labor-intensive services, and small-business-dependent industries, though the article does not specify which sectors beyond Barr’s general references. The central uncertainty remains whether AI evolves as a widely shared commodity or as a concentrated source of market power.