Are Robots Really Coming For Our Jobs?
What We’ve Learned From Previous Industrial Revolutions
By Arthur Daemmrich, Director of the Lemelson Center for the Study of Invention and Innovation, Smithsonian Institution, USA
A fourth industrial revolution?
According to a loose coalition of economists, techno-enthusiasts, and other analysts, we are in the early stages of a fourth revolutionary moment, during which innovations in technology bring about major changes to the production of goods and services, and significantly impact employment. The first such transformation came with the industrial revolution of the 1820s-1850s, the second through electrification and the innovation of mass production from 1900–1930, and the third with the introduction and spread of computers and information technology from 1950–1980.
In reports backed by the World Economic Forum, the World Bank, the United Nations, and numerous consulting firms, excitement about the capacity of new AI systems to manage complex tasks — ranging from long-distance trucking to operating the energy grid, to accounting and legal services — is tempered with warnings of sentient machines undercutting employment, data privacy, and even national security. Scholars, critics, and the public consequently have focused their attention on large complex technological systems and big data. But to understand the fourth industrial revolution only in terms of big technology would be to miss key insights about technology, employment, and shifting notions of intelligence that can be drawn from the history of technology.
Visible and invisible drivers of change
In 1776, Adam Smith began his famous text, An Inquiry into the Nature and Causes of the Wealth of Nations, with what sounded like an anecdote. Describing the making of metal pins used to shred and comb fibers, to tailor clothing, and to make combs and brushes, Smith noted that a layperson would struggle to make more than a single pin in a day. A skilled craftsman working alone could make 20 pins. But driven by the “invisible hand” of market forces, pin making was characterized by a division of labor: “One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations … to whiten the pins is another; it is even a trade by itself to put them into the paper [packaging].” Smith calculated that 10 workers could produce some 48,000 pins in a day, a major productivity gain over craftsmen working individually.
From Smith’s observation of pin making in the 1770s, two key points emerge at the center of contemporary discussions regarding technology, innovation, and employment. First, both visible and invisible forces are fundamental to the organization of work and the kinds of jobs available at a given historical moment. An “invisible hand” of market forces rewards efficiencies in production. For the individual, the invisible hand means they can build specialized skills and then sell or trade their output. For a small group, in this case the pin making company, new ways to increase production will help the firm prosper relative to competitors. For a larger collective — an economic region or even a nation — an invisible hand guides widespread adoption of specialized labor and greater total economic output using fewer labor hours. At the same time, visible managerial decisions also shape workplace structure. Initial equipment purchases that automate work and increase output relative to hours of human labor typically require significant up-front investments. Buying new technology is not so much driven by invisible market forces as by visible human decisions.
The controversial point — up to the present day — is whether the reorganization of production is driven by managerial choice or by invisible market forces. Will artificial intelligence be installed (at significant cost) in an effort to undermine knowledge workers — as machinery was used to displace skilled labor historically — or will it be so productive that market forces invisibly drive its widespread adoption?
A second key point arising from Smith’s work finds that as products and services become cheaper through the reorganization of work, consumers can buy more of them. A beneficial business cycle of greater production, greater employment (even as efficiency leads to less employment per unit of output), and total overall consumption results. In effect, lower-cost products can stimulate more consumption leading to more job growth. Of course, reduced demand from a general economic recession or from changes to consumer behavior also can lead producers to lower prices but without the benefit of greater demand. Observation of pin making in the 1770s nevertheless drew Smith to the conclusion that cheaper pins arose from the division of labor, and in turn enabled lower-cost production of cloth with significant benefits to broader society.
Several decades after Smith’s groundbreaking study, the industrial revolution would impact employment in ways unforeseen by Smith, but congruent with his theories. Breakthroughs in manufacturing technologies led to both job creation and job destruction, a pattern different from the past but now fundamental to all subsequent technology-based revolutionary periods.
History’s long tail
As the concept of a fourth industrial revolution takes hold and new policies are written for taxation, trade, and labor, it is instructive to explore the long tail of the first industrial revolution. We often look to technological history for straightforward stories of disruption. The average tenure of companies on the Fortune 500 list has declined from 67 years in 1925, to 33 years in 1965, to 20 years by 1995, to even less today.
Failure to deal with technological change, and to identify new markets, are given as leading causes for corporate failure in assessments by consulting firms. Likewise, situations in which one technology completely replaces another make for clean lessons about the failure to adapt to change. Carriage makers did not survive Ford’s mass-marketed automobile. Photos of New York City taken in 1910, 1920, and 1930 show a striking shift: from a streetscape dominated by horses and people walking, to one owned by cars. Pin making machinery, however, tells a more complex story. Innovations from the 1870s are still operating in at least one factory in the United States, even as sophisticated CNC machines have a 30-year lifetime, at most. Old and new can co-exist when supplying different submarkets.
Furthermore, the technological transitions in pin making are interesting in light of present-day concerns that the next wave of artificial intelligence will wipe out employment ranging from trucking to accounting. More realistically, innovations will help certain markets grow and will generate more employment overall, even as some specializations see jobs shrink or disappear. A fourth industrial revolution will also likely hide some of the hidden costs of lower-cost production of goods and services. We now know to look for them, but where? Could public policies be designed in advance, rather than only after today’s equivalent of rooms full of 7-to-10-year-old children, working long into the evenings, become visible?
The trope of “machines coming for your job” has become more widespread in the past five years than any time since at least the 1920s. It offers a self-fulfilling logic to those in the business of converting hype to funding to run their startups, namely that automation is inexorable and the main concern of public policy is to make the world safe for technology. Today’s artificial intelligence can only be thought of as such because we consider coding and sifting massive data sets to be markers of intelligence, just as metal work to make tiny pins was considered a skilled and intelligent craft until it was mechanized.
By looking to the pin instead of to the AI robot, we find an intriguing story of flexibility, adaptation, and innovation across periods of major technological, economic, and social change. From Adam Smith’s division of labor, to automation in the first industrial revolution, to supplying metal pieces machined to a third of the thickness of a sheet of office paper, a practice as simple as pin making is a stand-in for thousands of other production technologies that are easily overlooked, but that make up the core of any industrial revolution.
A new wave of AI equipment will support both additional change and constancy for companies. Today’s factories will take another evolutionary step in the now-long history of industrial evolution. As before, this will necessitate and reward new skills, but it will not eliminate work or employment.
A longer version of this article was originally published on Medium, on Mar 25, 2019. Read the original here.