Federal Reserve Governor Michael Barr has warned that the United States should prepare for serious short-term disruption to its labor market as artificial intelligence adoption accelerates, raising the possibility of widespread job displacement and a further shift in national income from workers towards investors and business owners.
Barr said the scale of the disruption would depend on how quickly companies deploy the technology and how widely its economic benefits are distributed. While AI could improve productivity and create new opportunities, he warned that rapid adoption could leave workers struggling to adjust, particularly younger employees and those with limited professional experience.
His comments, prepared for delivery at Yale University on Thursday, do not predict that mass unemployment is inevitable. Rather, they highlight the risks of a transition in which the technology advances faster than workers, businesses and institutions can adapt.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
The warning comes after US midterm elections in which public concerns about the rapid expansion of AI infrastructure contributed to a backlash against the technology buildout. It also precedes a public Federal Reserve forum in San Francisco on Monday, organized by a task force appointed by Fed Chairman Kevin Warsh to examine AI’s economic implications.
Barr said the long-term outcome would depend partly on whether competition among AI companies allows the technology to become widely accessible or whether economic gains remain concentrated among a small group of firms and investors.
Barr outlined two possible paths for the technology. In one scenario, competition among AI developers drives down costs and makes advanced systems widely available, allowing businesses of different sizes to benefit.
If AI becomes “a cheap and ubiquitous commodity, with gains widely shared,” he said, its impact could resemble the eventual spread of the internet, which expanded access to information and digital services.
In that environment, “AI could democratize access to skills and information needed for small businesses to thrive,” Barr said.
The alternative is a market dominated by a handful of companies that control the technology, computing infrastructure and investment needed to develop advanced AI systems. In that case, productivity gains and financial returns could accrue disproportionately to shareholders and owners of capital, while workers face pressure from automation.
Barr said some researchers believe AI has characteristics that, based on previous technological advances, could reinforce market concentration.
The concern is notably weighty given the decline in labor’s share of national income. This measure estimates how much of economic output goes to workers through wages and benefits, compared with the share accruing to businesses and investors through profits and other forms of capital income.
According to Barr, labor’s share fell to a record low of 52.8% in the second quarter of 2026. It had remained broadly between 60% and 65% from the Second World War until the early years of this century.
A further decline could mean that even as AI increases overall economic output, a smaller proportion of the gains reaches workers. The distribution of productivity improvements, rather than productivity growth alone, will therefore be central to determining the technology’s effect on household incomes and economic inequality.
“No one knows the degree to which the benefits will be shared,” Barr said.
He added that there had so far been relatively little evidence of broad worker displacement, with some indications that AI adoption has increased demand and created jobs. However, he said early signs of substitution were emerging in certain sectors, particularly among younger and less-experienced employees.
“If labor market changes happen quickly it will be hard for workers to adjust, and dislocations might be large,” Barr said. “The concentration of market power may help drive a widening of inequality that would strain the American Dream.”
The gap between job creation and job displacement is expected to widen as companies move from experimenting with AI tools to integrating them into core business operations. New demand for AI-related skills and services could generate employment, but that would not necessarily compensate workers whose existing roles are reduced or eliminated.
Barr’s warning also points to the risk that smaller businesses could struggle to compete if access to advanced AI systems remains expensive or concentrated among large technology companies. Their ability to use the technology to expand, rather than simply replace existing workers, could help determine how broadly its economic benefits spread.
Barr also questioned the assumption that AI-driven productivity gains will quickly reduce inflation, a possibility Warsh has frequently raised when discussing the technology’s long-term economic potential.
While AI could eventually allow companies to produce more with fewer resources, Barr said the immediate economic effects have been shaped by the enormous investment required to build the infrastructure supporting the technology.
AI companies are committing trillions of dollars to data centers, advanced processors and the computing resources needed to train and operate increasingly powerful models. That spending has contributed to US economic growth, Barr said, but has also increased demand for materials, equipment and skilled labor.
“The second big effect is on prices,” he said, adding that pressure from AI-related investment was spreading into other parts of the economy.
Barr cited higher costs for consumer products such as mobile phones and gaming consoles, as well as rising wages in the construction industry, where demand for workers has been supported by the expansion of data-center infrastructure.
The comments highlight a potential tension in the economic impact of AI. The investment required to build computing capacity can increase demand and inflationary pressure in the near term, even if the technology eventually improves productivity and reduces costs.
That timing matters for monetary policy. If infrastructure spending continues to support demand and push up input costs, the expected productivity benefits may not arrive quickly enough to ease inflation pressures.
Barr’s remarks also indicate that AI investment is becoming part of the Federal Reserve’s assessment of the inflation outlook. Several of his colleagues have acknowledged the role of technology-related spending in sustaining price pressures above the central bank’s target, strengthening the case among some policymakers for further interest rate increases.