The 1960s Automation Panic and Four Other Times Technology Was Supposed to Cause Mass Unemployment
In March 1964, a group of scientists, economists and activists sent President Lyndon Johnson a memo warning that machines were about to make most human work unnecessary. Less than two years later, unemployment fell below 4 percent.
It was not the first time Americans were told a new technology would cause mass unemployment, and it was not the last. Weavers feared it in the 1810s. Economists worried about it as machines spread through factories and tractors spread across farm country. Bank tellers were supposed to be finished by the ATM, and office clerks by the spreadsheet.
I went back through each scare, what was predicted at the time, and what actually happened, both to the jobs that were targeted and to employment overall.
The answer is not that everything worked out. Real people lost real livelihoods every time. But the mass unemployment never arrived, and the data shows why.
When America Was Sure Machines Would End Work
Five times a new technology was supposed to cause mass unemployment, with the 1960s panic at the centre. Each time, the job that was targeted shrank and the total number of jobs kept climbing.
- 1 1955: Congress holds hearings on automation
- 2 1964: The Triple Revolution memo
- 3 1969: First US ATM
- 4 1979: VisiCalc, the first spreadsheet
- 5 2022: ChatGPT released
Top: total nonfarm payroll jobs, annual average, 1939 to 2025. Bottom: the employment-population ratio, the share of Americans 16 and older with a job, 1948 to 2025. The 2025 share-of-adults figure is an 11-month average, because October 2025 was not surveyed during the government shutdown. Source: Bureau of Labor Statistics, via the Federal Reserve Bank of St. Louis.
Look at the top line and you might think the answer is obvious. The number of jobs in America went from 30.6 million in 1939 to 158.4 million in 2025, roughly 5 times as many. But most of that is population growth. More people means more jobs.
The bottom line is the real test. It shows the share of adults who actually have a job. If machines were steadily destroying work, that line would sink over time. It doesn't. It was 55.7% in 1964, the year of the memo, and it climbed to 64.4% by 2000, driven in large part by women moving into paid work. Its big drops line up with recessions, not with new inventions.
1964: the memo that said work was ending
The word itself was new. "Automation" was coined at the Ford Motor Company in the late 1940s, usually credited to vice president Delmar Harder, and popularised by John Diebold's 1952 book on the automatic factory. By the end of the 1950s it had become shorthand for a fear that factories, offices and mines were learning to run without people.
The fear grew with unemployment. The rate hit 7.1% in May 1961, and that February, Time magazine ran a story called "The Automation Jobless." It warned that what worried experts was not just the jobs being lost, but that "automation may prevent the economy from creating enough new jobs."
John F. Kennedy took it seriously. At a news conference on February 14, 1962, a reporter cited a Labor Department estimate that about 1.8 million workers a year were being replaced by machines. Kennedy answered:
Then came the memo. On March 22, 1964, a group calling itself the Ad Hoc Committee on the Triple Revolution sent Lyndon Johnson a document called The Triple Revolution. Its signers included the chemist Linus Pauling, a two-time Nobel laureate, the economist Gunnar Myrdal, the writer Michael Harrington, the activist Tom Hayden and, in the published version, the civil-rights organiser Bayard Rustin.
The memo described three revolutions happening at once: what it called the cybernation revolution, a revolution in weaponry and a revolution in human rights. The first one was the alarm. Cybernation, it said, was producing "a system of almost unlimited productive capacity which requires progressively less human labor," and "the traditional link between jobs and incomes is being broken." Its answer was a guaranteed income: "an unqualified commitment to provide every individual and every family with an adequate income as a matter of right."
The memo made the front page of the New York Times. The White House sent back a polite letter promising "thoughtful consideration." But Johnson was already moving. That August he signed the law creating the National Commission on Technology, Automation, and Economic Progress, saying at the signing: "Automation is not our enemy. Our enemies are ignorance, indifference, and inertia."
The commission, which included the economist Robert Solow, the union leader Walter Reuther and IBM's Thomas Watson Jr., reported in February 1966. Its central conclusion has been quoted ever since:
The commission found that the high unemployment of the late 1950s and early 1960s was not caused by machines at all. It blamed rising productivity and a growing labour force meeting "an inadequate response of aggregate demand." In other words, the economy wasn't spending enough to put everyone to work.
By the time the report came out, the problem was already solving itself. The 1964 tax cut was pumping money into the economy, and spending on the Vietnam War was adding more. Unemployment hit 4.0% in December 1965 and fell to 3.8% in February 1966. From then on, except for a single month at exactly 4.0%, it stayed below 4% for the rest of the decade, bottoming at 3.4%. The number of jobs grew from 58.4 million in 1964 to 70.5 million in 1969, and the share of adults working rose from 55.7% to 58.0%.
The losses were real, though. Coal is the clearest case. A 1961 Bureau of Labor Statistics study found that bituminous coal production jobs fell from about 411,000 in 1948 to under 150,000 in 1959, while output per worker nearly doubled. The commission itself recommended that the government act as an employer of last resort and put a floor under family incomes. Neither became law in the form proposed, though pieces of the idea showed up later in programs like Supplemental Security Income and the Earned Income Tax Credit.
The 1810s: the original Luddites
The first frames were smashed at Arnold, near Nottingham, on March 11, 1811. Over the next few years, groups of textile workers in the Midlands and the North broke stocking frames, shearing frames and power looms. Parliament's answer, in 1812, was to make breaking stocking frames a capital crime.
Lord Byron opposed the bill in his first speech in the House of Lords, on February 27, 1812. He described the problem plainly: "one man performed the work of many, and the superfluous labourers were thrown out of employment." The law passed anyway. The harshest punishment came the next year, though not under that law. At York in January 1813, more than 60 men were tried and 17 were hanged, for the murder of a mill owner, an armed attack on a mill and raids to seize weapons.
The Luddites were not against machines as such. The historian Eric Hobsbawm called their methods "collective bargaining by riot," a way of pressuring employers over wages and conditions. In Nottingham, the frames being smashed were the ones employers used to turn out cheap goods and to hire workers who hadn't served an apprenticeship. And the workers had reason to be afraid. According to economists Daron Acemoglu and Simon Johnson, hand-loom weavers in 1806 earned about twice what factory workers did. By 1820 they earned about a fifth less, and by the economists' estimate their real wages probably fell to around a quarter of their peak. Their numbers held up for years while their pay collapsed, rising from about 184,000 in 1806 to around 240,000 in the 1820s, before falling to about 23,000 by 1856.
But textile employment as a whole grew. Factory inspectors counted 423,400 workers in Britain's powered textile mills in 1838 and 682,497 in 1856. Cheaper cloth meant far more people could afford it, and the factories needed more hands to make it.
The 1930s: farms and "technological unemployment"
The biggest job loss in American history was on the farm. In 1900, about 41% of the American workforce worked in agriculture. By 1970 it was 4%, and by 2000 it was 1.9%. By USDA estimates, farm employment peaked at about 13.6 million people in 1916. By 1970 it was down to about 4.5 million.
Machines did most of it. The Census of Agriculture counted 246,083 tractors on American farms in 1920 and 3.6 million by 1950.
In the middle of that shift, during the Depression, the economist John Maynard Keynes gave the fear its most famous expression. American economists were already debating the phrase, but Keynes made it stick. In 1930 he wrote that the world was being afflicted with "a new disease":
Keynes himself called it a temporary phase of maladjustment, and he was right about that. The people who left farms, and their children, went into factories, offices, shops and schools. Nonfarm payroll jobs, the top line on the chart, more than doubled between 1939 and 1970, from 30.6 million to 71.0 million.
The 1970s: the ATM and the bank teller
The first American ATM went into service at a Chemical Bank branch in Rockville Centre, New York, on September 2, 1969. It was about as direct a replacement for a human job as a machine can be: it handed out cash so a teller didn't have to.
What happened next is the favourite example of economists who study automation. James Bessen, a lecturer at Boston University School of Law, found that the number of tellers needed to run a typical urban branch fell from 20 to 13 between 1988 and 2004. But that made branches cheaper to run, so banks opened more of them. Urban bank branches increased 43%. In Bessen's words, "the number of bank teller jobs did not decrease as the ATMs were rolled out." Bureau of Labor Statistics survey data put the number of tellers at about 288,000 in 1972, 531,000 in 1980 and 453,000 in 2010. Most of the growth came before ATMs were everywhere, but the job held up through the decades when they spread, as tellers shifted from counting cash toward sales and customer service.
It didn't last forever. Online and mobile banking did what the ATM couldn't. The Bureau of Labor Statistics counted 339,200 tellers in 2025, and it projects a further 13% decline over the following decade.
The 1980s: the spreadsheet and the bookkeeper
In June 1975, BusinessWeek ran a special report on "The Office of the Future." Four years later VisiCalc, the first electronic spreadsheet, arrived on the Apple II, and Lotus 1-2-3 followed on the IBM PC in January 1983. Work that took a bookkeeping clerk days could be done in minutes.
The clerks did lose out. According to an analysis of government jobs data by NPR's Planet Money, about 400,000 bookkeeping and accounting clerk jobs disappeared after 1980. But about 600,000 jobs for accountants were added over the same period, along with whole categories of financial analysts and office jobs that existed because the spreadsheet made it cheap to ask more questions of the numbers.
The honest caveat
Every one of these stories has the same two halves. The specific job that was targeted shrank, sometimes brutally. And the economy as a whole created more work than it destroyed. The people who lost out were rarely the ones who got the new jobs. A hand-loom weaver in 1820 didn't become a factory manager, and a laid-off coal miner in 1959 didn't become a computer programmer.
The research backs up the pain side. A study by economists Daron Acemoglu and Pascual Restrepo, published in 2020, found that across US local labour markets, "one more robot per thousand workers reduces the employment-to-population ratio by 0.2 percentage points and wages by 0.42%." The share of adults working peaked at 64.7% in April 2000 and had not returned to that level through 2025.
None of this proves what AI will do. Every past scare was about machines that replaced muscle or routine steps. Whether a technology that can do a wider range of thinking work follows the same pattern, or breaks it, is a genuinely open question. What the history does show is that the confident predictions of mass unemployment have been wrong every time so far, and that the real damage has landed on the specific workers in the way.
In 1964, some of the most respected thinkers in America told the President that machines were about to make most work unnecessary. Within two years unemployment was under 4%, and the national commission set up to study the problem concluded that technology eliminates jobs, not work.
The same pattern ran through the Luddites, the farm, the ATM and the spreadsheet: the targeted job shrank, total employment grew, and the cost fell hardest on the people doing the job that disappeared.
Sources: Bureau of Labor Statistics, Current Employment Statistics and Current Population Survey, via the Federal Reserve Bank of St. Louis (FRED); John F. Kennedy Presidential Library, news conference of February 14, 1962; The Triple Revolution (Ad Hoc Committee on the Triple Revolution, March 22, 1964); The American Presidency Project, remarks of August 19, 1964; Public Law 88-444; National Commission on Technology, Automation, and Economic Progress, Technology and the American Economy (1966); Economic Report of the President (1966); Time (February 24, 1961), as quoted by David Autor, Journal of Economic Perspectives (2015); BLS Bulletin 1305 (1961); Hansard, House of Lords, February 27, 1812; Eric Hobsbawm, The Machine Breakers (1952); Acemoglu and Johnson, Annual Review of Economics (2024); USDA Economic Research Service, The 20th Century Transformation of U.S. Agriculture and Farm Policy (2005); USDA farm employment series; Census of Agriculture (1954); John Maynard Keynes, Economic Possibilities for our Grandchildren (1930); James Bessen, Finance and Development (IMF, March 2015); BLS Occupational Outlook Handbook; NPR Planet Money; Acemoglu and Restrepo, Journal of Political Economy (2020).
Filed under: General Knowledge
