Federal Reserve governor Michael Barr describes a U.S. economy with a stabilizing labor market but very low job creation and an unemployment rate near its estimated long-run level, alongside personal consumption expenditures inflation at 3 percent and persistent risks of inflation remaining above the 2 percent target. He argues that monetary policy should remain cautious, with policy rates held steady until there is clear evidence of a sustained retreat in goods price inflation, before turning to how generative artificial intelligence appears increasingly likely to function as a general-purpose technology that reshapes production, business models, and innovation. Barr notes that artificial intelligence could also serve as an “invention in the method of invention” by accelerating research and development in areas such as drug discovery and materials science, but warns that the transition could be disruptive for workers in the short term.
Barr outlines three broad scenarios for the diffusion of artificial intelligence. In a gradual adoption path, artificial intelligence behaves like past general-purpose technologies, delivering stronger productivity growth, some occupation displacement, and time for workers and education systems to adjust, with real wages rising as the economy expands. In a more extreme “jobless boom” scenario, exponentially improving artificial intelligence and rapid adoption lead to widespread automation of professional, service, manufacturing, and transportation jobs, soaring layoffs, lower labor force participation, and a need to redesign the social safety net and workforce development. A third scenario envisions stalled advances and overinvestment, with artificial intelligence constrained by data, power, or capital, an “artificial intelligence bust” that delivers only modest and fading productivity gains, and a shift in risk from the labor market to the financial sector, echoing historical episodes of overbuilding such as railroads and fiber optics.
Early evidence, in Barr’s view, most closely resembles the gradual adoption scenario, with elevated productivity growth in recent years but only modest penetration of generative artificial intelligence so far and limited aggregate effects on employment or wages. As of December 2025, 17 percent of businesses in the Business Trends and Outlook Survey report using artificial intelligence, with about 30 percent of businesses with more than 250 employees reporting use, while a McKinsey survey finds that 88 percent of mostly large firms report using artificial intelligence in at least one business function and that the share using generative artificial intelligence specifically rose from 33 percent in 2023 to 79 percent in 2025. Research cited by Barr indicates that task-level access to artificial intelligence tools can raise efficiency and accuracy and could add between 0.3 and 0.9 of a percentage point to annual total factor productivity growth over the next decade, and internal Federal Reserve experiments show artificial intelligence tools cutting the time needed to update hundreds of databases by 50 percent and detecting and resolving 30 percent more issues during testing compared to earlier migrations. At the same time, studies suggest early-career workers in occupations highly exposed to artificial intelligence are already seeing weaker employment outcomes, while firms tend to reassign rather than lay off workers and plan to retrain staff, leaving long-run inequality and labor market effects highly dependent on policy choices in education, training, and social support.
Barr emphasizes that the distributional consequences of artificial intelligence are still uncertain, as artificial intelligence assistants may help less-experienced workers and reduce productivity gaps, but uptake so far is concentrated among younger, highly educated, high-income workers, which could widen wage inequality if those workers capture most of the gains. He stresses that artificial intelligence could alter demand across occupations, potentially automating higher-paying information jobs while increasing relative demand for lower-paying or less-educated roles, thereby reshaping but not necessarily increasing inequality, depending on ownership of artificial intelligence capital and complementary policies. For monetary policy, Barr warns that structural labor dislocation from artificial intelligence could raise the natural rate of unemployment in ways that interest rates cannot directly fix, while a sustained artificial intelligence-driven productivity boom could lift equilibrium interest rates, or r*, by boosting investment demand and expected lifetime earnings, and near-term artificial intelligence infrastructure spending could prove inflationary if energy supply lags demand. He concludes that artificial intelligence is likely to transform work and living standards, that the long-term benefits could be substantial, and that both the private and public sectors must act early and decisively to mitigate short-term worker dislocation and ensure the gains are broadly shared.
