TradingKey - On November 16, Eastern Time, Federal Reserve Chair Kevin Warsh said at the Productivity and Employment Working Group meeting in San Francisco that superintelligence (SI) could boost the long-term growth potential of the U.S. economy, but job displacement, investment returns failing to materialize, and falling stock valuations also need to be factored into policy considerations.
The question posed by Warsh's speech is: When technology begins to accelerate innovation itself, can productivity dividends translate into broad-based improvements in income and employment? The three economic scenarios he outlined illustrate different potential trajectories for growth, employment, and capital markets.
In Warsh's view, the difference between superintelligence and general-purpose technologies such as the internal combustion engine and electricity may lie in its ability to accelerate the generation of new ideas. While past inventions spurred subsequent innovation, superintelligence could also alter the methods and pace of innovation.
This assessment is grounded in endogenous growth theory: ideas are a vital resource driving long-term prosperity, and one person's use of a piece of knowledge does not diminish another's ability to use it. As the costs of acquiring specialized knowledge and exploring new questions fall rapidly, the fruits of innovation can spread more broadly, driving sustained economic growth.
However, technological potential does not yet equate to a realized productivity revolution. Warsh noted that while U.S. productivity has performed strongly in recent quarters, a clear structural upward breakout has yet to emerge. How large the technological gains will be, when they will begin, and how long they will last remain questions that warrant further study.
Warsh outlined three scenarios for the future, highlighting divergences between growth and employment. Warsh stressed that these are merely illustrative scenarios constructed by staff using models, rather than forecasts.
Scenario | Key Assumptions | Economic Growth |
Broadly Shared Productivity Dividends | Superintelligence boosts productivity, benefiting both workers and capital owners | Real GDP growth remains persistently above the long-term average |
Growth Accelerates but Jobs Are Displaced | Technology replaces some jobs, and barriers exist to labor reallocation | GDP growth is above the long-term average |
Technology Returns Fall Short of Expectations | Following substantial capital investments and stock market gains, expected returns decline | The economy slows, and upon eventual recovery, still fails to realize the expected productivity gains |
Warsh stated that the Federal Reserve's responsibility regarding superintelligence is to identify the economic impact of its widespread application and prepare for future monetary policy. Core variables the central bank needs to focus on include the scale, scope, and timing of productivity gains, as well as differences between the U.S. and other economies.
These scenarios indicate that assessing economic conditions in the future will require both observing whether growth accelerates and determining whether workers can share in the gains. Strong GDP and weak employment could coincide, which would make understanding economic performance more challenging.
He also warned that the vulnerability of critical infrastructure warrants serious attention. Regardless of how much knowledge machines acquire, human judgment, ethics, and responsibility remain irreplaceable, and final decisions must be made by humans.