Federal Reserve Chair Kevin Warsh said the central bank's staff have begun modeling how artificial intelligence could reshape the U.S. economy, including scenarios where automation raises unemployment despite economic growth.
Speaking at the Fed's Productivity and Jobs conference in San Francisco, Warsh described three possible outcomes over the next several years. He emphasized that these are illustrative scenarios rather than forecasts of what will actually occur.
Three Potential Paths
In the first scenario, AI increases productivity, allowing companies to produce more with the same resources. Economic output rises above historical growth rates, workers benefit from these gains, and unemployment remains near levels associated with a healthy labor market.
The second scenario also features faster economic growth, but businesses automate jobs more quickly than displaced workers can find new positions. Unemployment rises despite productivity improvements.
The third scenario assumes AI investments fail to deliver expected returns. Companies invest heavily in infrastructure, stock prices rise based on profit expectations, but those expectations prove too optimistic. Equity valuations could fall, leading to an economic slowdown and sharper unemployment increases before the economy recovers without the promised productivity gains.
Financial System Risks
Federal Reserve Governor Michael Barr raised separate concerns about how the AI infrastructure boom is being financed. The financing increasingly involves complex arrangements including long-term computing leases, energy contracts, and financial guarantees.
The Fed estimates that the maximum potential value of these off-balance-sheet obligations exceeds $3 trillion. While these commitments differ from debt already owed, disappointing AI revenues could create broader financial problems beyond technology companies building data centers.
Warsh did not propose interest-rate changes in his remarks. The Fed's next policy meeting is scheduled for December 8 and 9.


