AI Boom: Lessons from History

In a paper published in 1995, Timothy Bresnahan of Stanford University and Manuel Trajtenberg of Tel Aviv University set out what they saw as the characteristics of a general-purpose technology. It is supposed to be used in many industries, has an inherent capacity for continuous improvement and gives rise to “innovation complementarities” – that is, inspire knock-on innovation in the industries that use it. AI is getting widely adopted, getting better All day and is being deployed in more R&D contexts than ever before. So when does the economic revolution begin?

The first lesson of history is that even the most powerful new technology takes time to transform the economy. James Watt patented his steam engine in 1769, but steam power did not overtake water as a source of industrial horsepower until the 1830s in Britain and the 1860s in America. According to Nicholas Crafts of the University of Sussex, steam’s contribution to productivity growth in Britain peaked after 1850, nearly a century after Watt’s patent. In the case of electrification, major technological advances had been accomplished before 1880, but American productivity growth actually slowed from 1888 to 1907. Nearly three decades after the first silicon integrated circuit, Nobel-prize winning economist Robert Solow was still observing that the computer age could be seen everywhere but in productivity statistics. The computer-driven productivity boom didn’t occur in the US until the mid-1990s.

The gap between innovation and economic impact is partly due to fine-tuning. Early steam engines were wildly inefficient and consumed enormously expensive piles of coal. Likewise, the stellar performance of recent AI tools represents a vast improvement over those that fueled the boom of AI enthusiasm nearly a decade ago. (Siri, Apple’s virtual assistant, for example, was released in 2011.) A lack of capital could also slow deployment. Robert Allen of New York University Abu Dhabi argues that the sluggish growth in productivity growth in Britain’s industrialization reflected a lack of capital to manufacture plants and machines, which gradually went away as capitalists reinvested their fat profits. Invested.

More recent work has emphasized the time required to accumulate what is known as intangible capital, or the basic information needed to make effective use of new technology. Indeed, Erik Brynjolfsson of Stanford University, Daniel Rock of the Massachusetts Institute of Technology, and Chad Syverson of the University of Chicago suggest that a disruptive new technology may be associated with a “productivity J-curve”. Measured productivity growth can actually decline over the years. Or a new technology appears decades later, as firms and employees divert time and resources to studying the technology and designing business processes around it. Only later when these investments bear fruit does J move upwards. The authors acknowledge that AI-related investment capital in abstraction may already be underpinning productivity growth, although not by much yet.

Certainly for many, questions about AI’s impact on growth take a backseat to concerns about consequences for workers. Here the messages of history are mixed. There is good news: Despite the epochal technological and economic change, the fear of mass technological unemployment is felt like never before. Tech can and does take a toll on individual businesses, however, in ways that can prove socially disruptive. At the start of the Industrial Revolution, mechanization dramatically increased demand for relatively unskilled labor but crushed the earnings of artisans who had previously done much of the work, which is why some chose to join machine-smashing Luddite movements. chose. And in the 1980s and 1990s, the automation of routine work on factory floors and offices displaced many workers of modest means, while boosting employment for both high- and low-skilled workers.

yes very nice

AI can increase the productivity of workers of all different skill levels, even writers. Yet what this means for a business as a whole depends on whether the improved productivity and lower costs lead to a big jump in demand or only a modest jump. When the assembly line—a process innovation with GPT-like features—allowed Henry Ford to cut the cost of making cars, demand increased and workers benefited. For example, if AI increases productivity and lowers drug costs, it could lead to a huge demand for medical services and professionals.

There’s a chance that powerful AI will break the historical mold. A technology capable of handling almost any task that the typical individual can perform would usher humanity into uncharted economic territory. Yet even in such a scenario, the past holds some lessons. The economic growth sustained with the steam revolution, and the further boom that followed with electrification and other later innovations, was itself unprecedented. They inspired a tremendous scramble to invent new ideas and institutions to ensure that radical economic change translated into broad-based prosperity rather than chaos. It may soon be time to scramble once again.

Read more from our column on economics Free Exchange:

Have Economists Got Inflation Wrong? (26 January)

Could Europe End Up With Inflation Worse Than the US? (January 19)

A warning from history for a new era of industrial policy (January 11)

Sign up for more expert analysis of the biggest stories in economics, finance and the markets money TalksOur weekly subscriber-only newsletter.

© 2023, The Economist Newspaper Limited. All rights reserved. From The Economist, published under license. Original content can be found at www.economist.com