Federal Reserve Chair Kevin Warsh stated that superintelligence could accelerate idea generation and support long-term US economic growth, speaking at the Productivity and Employment Task Force Conference held in San Francisco earlier today. Warsh highlighted how advancements in technology might shape employment, outlining three distinct scenarios for the potential spread of this innovation across the economy.
Three scenarios for growth and unemployment
Researchers from the Federal Reserve Board have been modeling possible outcomes from the adoption of superintelligent technologies in the coming years. Warsh presented three examples based on these studies at the conference.
In the first scenario, increased productivity benefits both workers and capital holders equally. Under these conditions, the US economy grows at a pace above its long-term average, maintaining a near-full employment rate.
The second scenario also projects stronger-than-average growth. However, widespread automation eliminates certain jobs, while challenges in transitioning to new roles push unemployment rates above levels compatible with full employment.
The third scenario describes an initial surge in investment and stock market gains tied to new technologies, but anticipated returns from these innovations later diminish. Weaker equity valuations then slow the economy, and unemployment rises even more noticeably. Economic recovery occurs gradually, with expected productivity improvements from superintelligence failing to materialize.
| Scenario | Growth | Unemployment | Main Driver |
|---|---|---|---|
| 1 | Above average | Near full employment | Broad productivity gains |
| 2 | Above average | Above full employment | Automation transitions |
| 3 | Slows after initial surge | Higher, then recovers | Diminished returns |
Superintelligence and the pace of innovation
Drawing comparisons to innovations like electricity and the internal combustion engine, which enabled decades of further discoveries, Warsh argued that superintelligence may automate parts of creative processes and increase the speed at which new ideas emerge. He suggested that the marginal cost of solving highly specialized questions is declining rapidly and could approach zero over time.
Referring to the work of Nobel laureate economist Paul Romer, Warsh noted that when an idea is used by many, its economic value expands significantly.
Nevertheless, some analysts believe that US research efforts have increased more than twentyfold since the 1930s, even as research productivity has declined. Warsh described recent labor productivity gains as encouraging but stated that it remains uncertain whether these improvements are lasting.
Mini dictionary: Paul Romer is an American economist and Nobel laureate recognized for his research on endogenous growth theory, which explains how innovation and ideas are central to economic expansion.
Fed examines policy implications
The Productivity and Employment Task Force, established by Warsh, is one of five working groups he leads at the Federal Reserve. Warsh explained that the Fed is assessing the extent to which technology will enhance the productivity of labor and capital, the likely timeline for these shifts, and how the broader diffusion of expertise may shape living standards.
He also warned that superintelligence could boost the capacity of malicious actors, emphasizing the importance of protecting critical infrastructure. Warsh added that as technology becomes more widespread, humans must retain responsibility for decision-making.
“Humans design and build. People must also make the decisions.”
From a monetary policy perspective, a key focus for the Fed will be how quickly new jobs replace those lost to automation. If superintelligence increases output but the workforce transitions slowly into new roles, the central bank may have to navigate a period of high growth alongside rising unemployment.






