Recent work by Korinek, Jones, Sacher, Cotter, and McCrory develops a transparent framework for assessing the economic consequences of transformative artificial intelligence through 2030. Their scenarios span an unusually wide range: relative to a no-AI baseline, U.S. GDP in 2030 is 1.6 percent higher in the modest scenario, 8.3 percent higher in the substantial scenario, and 32.4 percent higher in the extreme scenario. The latter implies annual growth rates with essentially no peacetime historical analogue and a substantial redistribution of income from labor toward capital.

 

This paper asks what these technological scenarios would imply once embedded in a forward-looking macroeconomic environment. Transformative AI would affect the economy not only through task productivity and labor displacement, but also through expectations, investment, capital accumulation, asset prices, interest rates, fiscal positions, international spillovers, and policy responses. These feedbacks may begin well before the underlying technologies are fully deployed.

 

We therefore propose a broader macroeconomic research and policy agenda that translates alternative AI futures into internally coherent transition paths and evaluates monetary, fiscal, financial, structural, and distributional policies under deep uncertainty. The central implication is that transformative AI should be treated not as a single productivity shock, but as a potentially persistent structural transformation whose macroeconomic consequences depend critically on expectations, adjustment dynamics, and policy.