The rapid expansion of artificial intelligence could create a sharp imbalance in the US economy, allowing companies and investors that own the technology to capture a growing share of national income while workers face job losses, weaker bargaining power, and limited opportunities to adjust, Federal Reserve Governor Michael Barr has warned.
Barr said on Thursday that the United States should prepare for “serious short-term disruptions” to employment as AI adoption accelerates, particularly if businesses deploy the technology faster than workers and institutions can adapt. He cautioned that widespread displacement could deepen inequality by shifting economic gains away from labor and towards owners of the capital and companies driving the AI boom.
His remarks, prepared for delivery at Yale University, did not amount to a prediction of mass unemployment. Instead, Barr stressed that the economic consequences remain uncertain and will depend on how quickly AI spreads, how widely its benefits are distributed and whether competition prevents a small group of technology companies from controlling the market.
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 amid growing political concern over the scale of AI investment and its potential consequences for employment. The recent US midterm elections were shaped partly by a backlash against the rapid expansion of AI infrastructure. Barr’s comments also precede a public Federal Reserve forum in San Francisco on Monday, where a task force appointed by Fed Chairman Kevin Warsh is expected to examine the technology’s implications for the economy.
The debate is becoming more consequential for policymakers because AI presents two potentially conflicting outcomes. It could raise productivity, lower business costs and make advanced capabilities accessible to smaller companies. But it could also concentrate wealth and market power among a handful of firms while reducing the share of economic output paid to workers.
Barr identified competition as a major factor that could determine how the benefits of AI are distributed. If more companies develop competing systems and technology becomes a cheap, widely available commodity, businesses of different sizes could gain access to tools that improve productivity and expand economic activity.
In that scenario, AI could lower barriers to entry for smaller businesses by giving them access to skills, information, and capabilities that were previously expensive or difficult to obtain.
“AI could democratize access to skills and information needed for small businesses to thrive,” Barr said.
However, he warned that a different outcome could emerge if control of AI systems and the infrastructure supporting them becomes concentrated among a small number of companies and investors. Those firms could capture a disproportionate share of the productivity gains, while workers whose tasks are automated face pressure on wages and employment.
Barr pointed to the declining share of national income going to labor as an indication of the distributional challenge. The labor share, which measures the portion of economic output paid to workers rather than accruing to business owners and investors, fell to a record low of 52.8% in the second quarter of 2026.
That compares with a range of roughly 60% to 65% maintained from the Second World War until the beginning of this century.
The decline matters because economic growth does not automatically translate into higher wages or improved living standards for workers. If productivity increases while a growing portion of the additional income accrues to shareholders and owners of technology, the economy could become more productive without distributing the gains broadly.
“Some researchers believe that AI has characteristics that have, in past example of high-tech advances, reinforced concentration of market power,” Barr said.
The concern is needs urgent attention because developing and operating advanced AI systems requires substantial computing infrastructure, investment and technical resources. Companies able to finance large-scale data centers and secure access to specialized chips may have advantages that smaller competitors struggle to match.
Barr said there had been relatively little evidence of large-scale worker displacement so far. Some businesses have also used AI to increase demand and create jobs. But he cautioned that early employment trends do not settle the question of what happens as the technology becomes more capable and more widely adopted.
He identified younger and less-experienced workers as a group already showing signs of exposure to AI-driven labor substitution. Entry-level roles often involve tasks that can be automated or assisted by AI systems, potentially reducing opportunities for workers to gain experience and progress into more senior positions.
“If labor market changes happen quickly it will be hard for workers to adjust, and dislocations might be large,” Barr said, adding that concentrated market power could widen inequality and strain the American Dream.
The speed of adoption is therefore as important as the technology’s ultimate capabilities. Gradual changes may give workers time to retrain, move between industries, or develop complementary skills. Rapid deployment across multiple sectors could leave employees with fewer alternatives, particularly in regions where employment depends heavily on industries undergoing automation.
AI Infrastructure Spending is Adding to Inflation Pressures
Barr’s warning also raises questions about the Federal Reserve’s assessment of AI’s economic effects. Warsh has frequently discussed the possibility that AI could usher in a period of stronger productivity growth, allowing businesses to produce more with fewer resources and potentially reducing inflation over time.
Barr said the timing of those productivity gains remains uncertain. So far, he argued, one of the clearest economic effects of the AI boom has been the inflationary pressure generated by enormous spending on data centers, computing equipment and the resources needed to operate them.
AI investment has supported US gross domestic product, but the demand for infrastructure has also increased costs across parts of the economy.
“The second big effect is on prices,” Barr said, describing how input pressures from the AI buildout were spreading into other industries.
He cited higher prices for consumer products such as mobile phones and gaming consoles, alongside rising wages in the construction industry as data center developers compete for workers and other resources.
This poses a difficult policy challenge. AI could eventually reduce production costs and ease inflation, but the investment required to build the infrastructure may increase demand for equipment, labor and materials in the near term. The benefits may take years to emerge, while the costs are already feeding through supply chains.
The inflationary impact of AI investment has gained attention among some Fed policymakers and has strengthened support for further interest rate increases. Higher borrowing costs could, in turn, affect the financing of data centers and other capital-intensive projects, while also weighing on households and businesses.
The broader issue for the Fed is that AI may affect the economy through different channels at different speeds. Infrastructure spending can raise output and prices before productivity improvements become widespread. Later, automation and efficiency gains could change labor demand, wage growth and the distribution of income.
Barr’s remarks indicate that policymakers cannot assess AI solely through its potential to lift productivity. They must also consider who owns the technology, who benefits from its deployment, and how quickly workers can adapt when businesses reorganize around it.