The conference was organised by the Federal Reserve Board and the San Francisco Fed.
Three paths, no forecast
Warsh described the scenarios as neither forecasts nor predictions. Fed staff have begun evaluating each, he said. One path sees AI lift productivity broadly. A third involves a reversal in equity valuations alongside a broader slowdown.
“Expected returns could reverse, equity valuations could fall, and this could lead to an economic slowdown and a more pronounced rise in unemployment,” Warsh said. “Over time the economy would recover, but without productivity gains.” He said research activity has grown substantially since the 1930s, and that a hypothetical doubling of ideas and other basic inputs could raise economic output by more than two times.
Neutral stance on technology
Warsh said productivity growth has been strong in recent quarters, though it has not yet settled into a lasting upward trend. He was careful to frame the Fed’s posture as neutral toward the technology itself.
“Our role at the Fed is neither to be supportive nor opposed,” Warsh said. “Our responsibility is to recognise that the broad adoption of this technology will have a very important effect on the economy and extremely important consequences for monetary policy.” He added: “Our job is to prepare for what is next.”
The US unemployment rate stood at 4.1 percent in October, its lowest in 31 days. Warsh moderated the conference’s first panel after delivering his remarks.
Data: FRED — St. Louis Fed








