Generative AI may yet live up to the hype. Such is the cautiously hopeful verdict from a new Federal Reserve discussion paper evaluating whether artificial intelligence models like ChatGPT are economic fads or foundational breakthroughs. Drawing on economic theory, historical precedent, and early productivity studies, the authors ask: Is GenAI a light bulb, a dynamo, or a microscope?
That clever framing comes from “Generative AI at the Crossroads: Light Bulb, Dynamo, or Microscope?” by Martin Neil Baily and Aidan T. Kane of the Brookings Institution, and David M. Byrne and Paul E. Soto of the Federal Reserve Board. In the light bulb scenario, GenAI delivers a one-time productivity boost that raises output levels but not long-run growth. Such inventions “temporarily raise productivity growth as adoption spreads, but the effect fades when the market is saturated; that is, the level of output per hour is permanently higher but the growth rate is not.”
So a spark, not a sustained surge. In the more optimistic dynamo and microscope models, GenAI becomes either a general-purpose technology that drives complementary innovation across the economy (“new goods and services, process efficiencies, and business reorganization”) or a powerful research tool that transforms the innovation process itself. Preferably both.
This dual potential is what makes GenAI so exciting. Technologies like the electric dynamo rewired how production was organized, eventually delivering sweeping, long-term gains. Likewise, the microscope helped humanity understand the hidden structures of the world, accelerating science itself. Gen AI seems poised to accelerate how we work and how we invent, “an encouraging sign that GenAI will raise the level of productivity.”
The latter capability, AI as an “invention of a method of invention,” is especially important because it could reverse a critical but underappreciated crisis: Innovation is becoming exponentially more expensive and difficult across industries. AI promises to rescue innovation from its cost spiral through three mechanisms, according to the McKinsey report, “The next innovation revolution—powered by AI.”
First, it can generate design candidates with unprecedented speed and creativity—conjuring thousands of molecular structures or engineering configurations that human researchers might never conceive. Second, AI surrogate models can replace expensive physical testing, simulating aerodynamics or drug interactions in minutes rather than months. Third, it streamlines the tedious work of synthesising vast literature and documentation that clogs modern R&D.
The potential rewards are substantial: McKinsey reckons this could unlock $360–560 billion annually, doubling pharmaceutical R&D productivity and boosting manufacturing by half.
As I write in my 2023 book, The Conservative Futurist: How to Create the Sci-Fi World We Were Promised, an invention in the method of invention “can have a permanent impact on how humans discover knowledge and solve problems across all fields of science research and technological development.”
Yes, GenAI adoption remains uneven, especially outside large tech-savvy firms. But that’s no red flag. The electric dynamo took decades to reshape industry, with impact accelerating only after workflows and factory layouts were redesigned around it. GenAI’s path is likely similar, at least directionally: not viral disruption, but patient restructuring.
If the trends hold, GenAI could do more than help us work faster at what we already do. It could help us discover and use new things. In the long run, that may be its most transformative contribution.
source: American Enterprise Institute
