Artificial Intelligence has become the defining technological debate of our era. Across universities, think tanks, media platforms, and international institutions, a growing number of prominent scholars warn that AI will produce mass unemployment, deepen inequality, and destabilise modern capitalism. Some even describe AI as an existential threat to the labour market itself. Interestingly, many of these arguments increasingly resemble the ideas of Karl Marx.
More than 150 years ago, Marx argued that capitalism’s constant drive toward innovation and mechanisation would eventually replace workers with machines. According to Marx, technological progress under capitalism would create a ‘reserve army’ of unemployed labourers, weaken workers’ bargaining power, and concentrate wealth into the hands of capital owners.
Today’s AI debate often echoes the same logic. The fear is no longer limited to factory workers losing jobs to machines. AI now threatens white-collar and knowledge-based professions once considered immune to automation – including accountants, lawyers, programmers, journalists, analysts, and even university educators. Yet despite the renewed popularity of technological pessimism, history tells a very different story.
Since the Industrial Revolution, technological innovation has repeatedly displaced certain forms of labour while simultaneously creating entirely new industries, occupations, and economic opportunities. Textile machinery replaced hand weavers, tractors reduced agricultural labour, and computers automated clerical work. However, none of these revolutions created permanent mass unemployment. Instead, economies adapted. Industrialisation generated manufacturing industries. Electrification created modern infrastructure and services. The computer revolution gave birth to software engineering, digital communications, cybersecurity, and the internet economy. Human labour did not disappear. It evolved.
This historical reality does not mean that AI poses no challenges. It certainly does. But the primary challenge is likely to be economic polarisation rather than universal unemployment.
The modern labour market is increasingly divided into two broad groups. On one side are highly skilled workers capable of using AI to dramatically increase productivity and income. On the other side are lower-wage service workers whose jobs remain difficult to automate completely. Between these two groups lies a shrinking middle class composed largely of routine administrative and repetitive professional occupations vulnerable to automation. In this sense, AI may weaken the middle layers of the economy while strengthening both the high-productivity elite and low-wage service sectors.
However, one important aspect often ignored in the current debate is that AI may actually increase the economic power of qualified individuals rather than simply strengthening corporations alone.
Marx argued that capital owners capture the ‘surplus value’ generated by workers. But the AI economy may partially alter this relationship. Today, a skilled individual equipped with AI tools can perform tasks that previously required entire teams or organisations. A software engineer can develop products faster. A researcher can analyse larger datasets. A consultant can serve global clients more efficiently. A content creator can automate production and distribution. In other words, AI may allow qualified labour itself to capture more surplus value.
This is one of the most significant transformations of the knowledge economy. Productive capacity is no longer confined entirely within large industrial institutions. AI increasingly amplifies individual capability.
Another overlooked issue is working hours. Marx believed industrial capitalism would continuously intensify labour exploitation. During the early factory era, working conditions indeed became harsh and exhausting. But over the long run, the opposite trend emerged in many advanced economies. As productivity increased, average working hours gradually declined while standards of living improved. AI may accelerate this process further.
If productivity rises substantially through AI systems, societies may eventually move toward shorter workweeks, greater flexibility, and reduced dependence on traditional full-time labour structures. The future may involve less work rather than no work.
Equally important is the historical rise in minimum welfare standards. Since Marx’s time, humanity has experienced extraordinary improvements in living conditions. Global poverty rates have declined dramatically. Life expectancy has doubled. Literacy, healthcare, transportation, communication, and access to information have expanded across the world. Even many lower-income households today possess technological conveniences unimaginable to nineteenth-century elites.
This does not mean inequality has disappeared. Wealth concentration remains a major issue and may intensify further in the AI era. But rising inequality and living standards can coexist simultaneously. The evidence suggests that technological advancement has historically increased overall human prosperity, even when wealth distribution remained uneven.
Most major declines in welfare have emerged not from technological innovation itself, but from war, political instability, corruption, institutional collapse, and economic mismanagement.
The current AI debate, therefore, requires balance and historical perspective. AI will undoubtedly transform labour markets. Some occupations will disappear. Others will emerge. The middle class may face increasing pressure. Inequality may rise if societies fail to invest in education, adaptation, and institutional reform. But predictions of inevitable mass unemployment repeat fears that accompanied nearly every previous technological revolution.
History suggests a more nuanced outcome: societies become wealthier overall, productivity rises, working hours decline, and new forms of employment emerge – even as economic structures change profoundly.
The central challenge of the AI age is not whether humanity will run out of work. It is whether societies can ensure that the benefits of AI remain broadly distributed rather than concentrated within a narrow technological elite. That question is political, educational, and institutional – not technological alone.
The writer is a Senior Associate Professor, Blekinge Institute of Technology, Sweden
