What the Luddites understood about artificial intelligence

Britain once made breaking a machine punishable by death.

At the height of the Industrial Revolution, Parliament's Frame Breaking Act of 1812 made the destruction of certain textile machinery a capital offense, concluding that attacks on the country's expanding manufacturing industry warranted the harshest punishment available under the law.

The law targeted the Luddites, groups of skilled textile workers who had begun smashing machinery across England's manufacturing districts the previous year. History remembers them as enemies of technological progress. In reality, they were protesting something far more specific: factory owners using new machinery to replace skilled labor, suppress wages and shift power from workers to employers.

More than two centuries later, artificial intelligence has revived many of the same questions. As companies and governments race to deploy generative AI, workers and policymakers are again debating who benefits from technological change, who bears its costs and who controls its deployment.

A rebellion over the terms of progress

Luddism emerged during a period of war, high food prices and severe economic pressure in the English textile industry. Beginning in Nottinghamshire in 1811 and spreading into Yorkshire and Lancashire, workers attacked machinery they believed was being used to undercut wages, evade apprenticeship rules and weaken established production standards.

Many Luddites were highly skilled textile craftsmen, including croppers and weavers.  Their letters and petitions did not reject technology itself, but objected to manufacturers using machinery to replace skilled artisans with lower-paid labor.

Mechanization initially benefited many textile workers. As spinning machinery increased the supply of yarn, the number of Britain's hand-loom weavers rose from about 37,000 in 1780 to 208,000 in 1812. Once power looms became commercially viable, manufacturers mechanized weaving and moved production into factories.

Workers attempted to secure protection through political means. A petition signed by approximately 130,000 people sought a legislated minimum wage for weavers, but Parliament rejected the proposal in 1808. Four years later, lawmakers passed the Frame-Breaking Act and made the destruction of specified machines a capital crime.

The legislation was followed by a sweeping crackdown. The government deployed thousands of troops into the affected regions, while suspected Luddites were imprisoned, transported or executed. Fourteen men were hanged in York on Jan. 16, 1813. The state had rejected many demands to regulate how machinery affected labor, but it imposed its harshest punishment to protect industrial property.

Marx and the machinery of capital

The conflict between machinery and labor later became central to the work of Karl Marx, the German philosopher, economist and revolutionary socialist whose critique of capitalism helped lay the intellectual foundations of modern communism.

Marx explored the issue in Chapter 15 of the first volume of Capital, published in 1867. In "Machinery and Modern Industry," he argued that machines could increase productivity and reduce labor, but under capitalism they were often introduced to increase what he called surplus value. Employers could use machinery to intensify production, replace skilled workers with cheaper labor and capture a greater share of the wealth workers created.

For Marx, the problem was not machinery itself but ownership and power.A technology capable of reducing physical effort could instead increase the pace of work or weaken workers’ position if its use was controlled entirely by employers.

Marx gave philosophical expression to what the Luddites had confronted decades earlier. Their protest was never about whether machinery could produce more cloth, but whether higher productivity would come at the expense of wages, autonomy and skilled work.

Keynes and the prospect of technological unemployment

British economist John Maynard Keynes revisited the issue in 1930 in Economic Possibilities for Our Grandchildren, introducing the term "technological unemployment" to describe labor displaced faster than new work could be created.

In his 1930 essay Economic Possibilities for Our Grandchildren, written at the beginning of the Great Depression, Keynes introduced the term “technological unemployment.” He defined it as unemployment caused when societies discovered ways to economize on labor faster than they could find new uses for displaced workers.

Keynes believed technological progress would ultimately make societies far wealthier and even predicted a 15-hour workweek. But he warned that the transition could impose severe hardship because higher productivity did not automatically translate into greater economic security

His optimism came with an important qualification. Technological unemployment could impose serious hardship during the transition, even if productivity eventually created abundance. The capacity to produce more with less labor did not itself determine whether people would receive higher incomes, shorter working hours or greater economic security.

That problem resurfaced during the automation debates of the 1960s, as computers, industrial robots and numerical control systems entered factories and offices. Each wave of innovation raised predictions of mass unemployment. New occupations eventually appeared, but the gains were never distributed automatically or evenly across workers, industries and countries.

AI changes work before it eliminates jobs

Research on AI points to a similar distinction between technological capacity and economic outcome. Rather than eliminating entire occupations, generative AI is expected to automate some tasks, augment others and alter the authority workers exercise over their daily work.

The International Labour Organization's March 2026 working paper Disruption Without Dividend? examined 135 countries representing 69% of the world's population. It found that generative AI's impact depends not only on occupational exposure but also on digital infrastructure, computing resources and workforce skills, leaving many developing economies vulnerable to disruption without comparable productivity gains.

AI is also extending automation into management. The OECD's December 2025 report How Widespread Is Algorithmic Management in Workplaces? found that at least one algorithmic management tool was used by 90% of surveyed firms in the United States, an average of 79% across France, Germany, Italy and Spain, and 40% in Japan.

Among surveyed European firms, 69% used such tools to instruct workers, 67% used them to monitor workers and 35% used them to evaluate performance. These systems can allocate schedules, track work speed, analyze communications, establish targets and recommend rewards or sanctions.

The original Luddites understood that a machine’s economic consequences could not be separated from the people who owned it and the rules governing its use. Artificial intelligence may generate major productivity gains, but whether they produce higher wages, shorter working hours or greater inequality will depend on ownership, regulation and worker participation. The machinery has changed. The struggle over who benefits from it has not.