When Capitalism No Longer Needs Workers

Artificial Intelligence and the Future of the Global Economy

Anxiety about the future of work has become a defining feature of the modern age. Artificial intelligence sits at the centre of that unease, occupying an uncomfortable space between dystopian fears of total automation and optimistic visions of a technologically enabled utopia. For most people, however, the concern is far more immediate and personal: the possibility that machines will soon perform their jobs faster, cheaper and better than they ever could.

Historically, societies have faced similar moments of uncertainty. Major technological breakthroughs—from the steam engine to electricity, from mechanisation to computers—have repeatedly transformed economies. Each time, jobs were displaced, but new industries eventually emerged. Productivity rose, wealth expanded, and employment shifted rather than vanished. This historical pattern has led many economists and policymakers to argue that artificial intelligence will follow the same trajectory.

Others are far less convinced. Previous waves of automation largely replaced physical labour, freeing humans to focus on cognitive and creative tasks. Artificial intelligence, by contrast, is increasingly capable of replicating precisely those mental skills. If machines can reason, communicate, analyse and create, what unique economic value remains for human workers? This question no longer belongs to the distant future. In several parts of the world, its consequences are already unfolding.

What makes the current transition particularly significant is not only the scale of potential disruption, but also its uneven distribution. Artificial intelligence is reshaping global capitalism in ways that threaten to widen the gap between rich and poor countries, between highly skilled and routine workers, and between those who own capital and those who rely on wages. To understand what happens when capitalism no longer needs workers in the way it once did, it is necessary to examine where the impacts are being felt first, why they are occurring, and what options remain for adaptation.

The Front Line of Automation in Developing Economies

While debates about artificial intelligence often focus on advanced economies, some of the most immediate risks are concentrated elsewhere. Over the past three decades, many developing countries have built growth strategies around outsourced service work. Nations such as the Philippines and Bangladesh invested heavily in education, language skills and digital infrastructure to attract business process outsourcing. Call centres, data entry, transcription services and basic software support became engines of employment and sources of foreign income.

These industries were long considered relatively safe from automation. They required natural language fluency, contextual understanding and interpersonal skills—qualities that machines struggled to replicate. That assumption no longer holds. Advances in large language models and related technologies have enabled software to perform many of these tasks with remarkable speed and accuracy, often at a fraction of the cost of human labour.

In the Philippines, outsourcing has grown into a multi-billion-pound sector employing more than a million people and contributing a substantial share of national output. Yet estimates suggest that nearly nine out of ten outsourced service roles in the country face a high risk of automation. For a workforce built around precisely the types of tasks artificial intelligence excels at—repetitive, rules-based and text-driven—the threat is existential.

Bangladesh faces a similar challenge, albeit on a smaller scale. Its outsourcing industry has expanded rapidly, employing tens of thousands of workers across hundreds of firms. Much of this work also centres on customer service, transcription and data processing. As artificial intelligence becomes cheaper and more capable, the economic logic that once favoured offshore labour weakens. If the same output can be delivered faster, more consistently and without the costs associated with human employment, companies have little incentive to continue outsourcing.

This shift undermines an entire development model. For decades, lower-income countries sought to integrate into the global economy by providing labour-intensive services to wealthier nations. Artificial intelligence threatens to reverse that flow. Tasks once sent abroad may instead be automated locally, or eliminated altogether. In extreme cases, a single server rack could replace an entire call centre, erasing thousands of jobs overnight.

The result is a harsh irony. Just as some developing economies began to close the gap with richer nations, the technological foundations of their progress are being destabilised. Artificial intelligence, rather than accelerating convergence, risks entrenching divergence.

Productivity Gains and Unequal Rewards

Artificial intelligence is undeniably capable of boosting productivity. Studies consistently project significant economic gains in countries that adopt these technologies at scale. Advanced economies with strong digital infrastructure, high levels of education and deep capital markets are expected to see substantial increases in output over the coming decade.

Lower-income countries, by contrast, are projected to experience more modest gains. This represents a departure from the long-standing assumption that developing economies grow faster by adopting existing technologies. In the case of artificial intelligence, the benefits appear to be disproportionately captured by those who already possess the necessary resources to deploy it effectively.

Several structural factors explain this imbalance. Developing, training and deploying advanced AI systems requires significant investment, reliable infrastructure and specialised expertise. These inputs are overwhelmingly concentrated in wealthy nations. As a result, the most valuable jobs associated with artificial intelligence—researchers, engineers, system architects and product designers—are largely inaccessible to workers in poorer regions.

Even when individuals from developing countries acquire these skills, they are often absorbed into global technology hubs. Talented engineers and researchers migrate to established centres where opportunities, funding and professional networks are more abundant. This brain drain deprives their home countries of precisely the human capital needed to build competitive AI ecosystems, further weakening their position.

Meanwhile, companies that control artificial intelligence technologies benefit from powerful network effects. The more data they collect, the better their systems perform. Better performance attracts more users, generating more data in a self-reinforcing cycle. This dynamic leads to increasing concentration of market power and profits in a small number of firms, most of them based in a handful of countries.

Estimates of artificial intelligence’s contribution to global output run into the tens of trillions of pounds over the coming years. Yet the majority of that wealth is expected to accrue to just two countries. Ownership of AI, rather than its use, is what determines who benefits most. For economies positioned primarily as consumers of AI rather than creators of it, the returns are far smaller.

Complementary and Substitutive Capital

To understand how artificial intelligence reshapes labour markets, it is useful to distinguish between two types of technological capital. Historically, many innovations functioned as complementary capital: they enhanced human productivity without eliminating the need for workers. Mechanised farming equipment, for example, allowed fewer people to cultivate larger areas of land, but human labour remained essential. As productivity increased, wages and living standards tended to rise alongside it.

Artificial intelligence behaves differently depending on the nature of the work involved. In high-skilled roles that require judgement, creativity and accountability, AI often acts as a complement. Professionals who can effectively integrate these tools into their work become more productive and more valuable. A financial analyst supported by advanced data analysis software can identify patterns more quickly and focus on strategic decisions. A doctor using AI-assisted diagnostics can spend more time on patient care.

In such cases, artificial intelligence amplifies human capability rather than replacing it. Demand for skilled labour may even increase, particularly for roles that combine technical proficiency with domain expertise.

For routine, process-driven tasks, however, artificial intelligence increasingly functions as substitutive capital. Rather than making workers faster, it renders them unnecessary. An automated customer service system does not assist a call centre agent; it replaces them. A sophisticated code-generating system does not support a junior developer; it eliminates the role altogether.

The more capable artificial intelligence becomes, the less dependent it is on human input. In this sense, capitalism is approaching a novel configuration: one in which productivity growth no longer requires proportional increases in employment. This represents a fundamental break from the economic dynamics that underpinned industrial capitalism.

Compounding this shift is the concentration of capital ownership. The assets that power artificial intelligence—data centres, proprietary algorithms, patents and specialised hardware—are owned by a relatively small group of firms and investors. As labour becomes less central to production, returns increasingly flow to those who own these assets rather than those who sell their labour.

Historical Parallels and Persistent Scars

Technological disruption is not new, and history offers sobering lessons about its social consequences. The automation and globalisation waves of the late twentieth century transformed manufacturing in advanced economies. Millions of factory jobs disappeared, particularly in regions heavily dependent on industrial employment.

In the United States, the decline of manufacturing devastated entire communities. Cities once defined by steel production, car manufacturing and textiles experienced factory closures, rising unemployment and long-term economic stagnation. The loss of stable, well-paid jobs reverberated through local economies, affecting everything from public health to educational outcomes.

The United Kingdom underwent similar transformations. The decline of coal mining, shipbuilding and steel production reshaped large parts of Northern England, Scotland and Wales. Although national output eventually recovered, regional inequalities persisted. Decades later, many former industrial areas continue to lag behind in income, employment and social mobility.

These experiences highlight a critical point: even when technological change leads to long-term economic gains, the transition can inflict lasting damage. Once inequality becomes embedded, it is extraordinarily difficult to reverse. Communities that lose their economic base often struggle to attract new industries, and displaced workers face barriers to retraining and re-employment.

Artificial intelligence threatens to replicate these patterns on a much larger scale, affecting not just specific regions or sectors but entire categories of work across the global economy.

Diverging Futures Within Countries

The impact of artificial intelligence is not limited to international disparities. Within individual countries, it is creating sharp divisions between workers. Those who can leverage AI effectively see their productivity and earnings rise. Those whose roles are easily automated face job loss or downward pressure on wages.

Surveys suggest that a significant proportion of workers are already worried about automation. This anxiety reflects a growing awareness that the changes underway are not hypothetical. In many sectors, artificial intelligence systems are already being deployed to cut costs and increase efficiency. Employers have strong financial incentives to replace labour where possible, particularly after investing heavily in new technologies.

This dynamic risks hollowing out labour markets. If businesses can grow output without employing more people, productivity may rise while consumer demand stagnates. Workers without income cannot spend, undermining the very markets that drive economic growth. Over time, this imbalance can become a drag on innovation and stability.

Social safety nets play a crucial role in mitigating these risks. Unemployment benefits, retraining programmes and income support can provide displaced workers with the time and resources needed to adapt. Without such measures, technological progress may come at the cost of widespread insecurity and social fragmentation.

Policy Responses and the Challenge of Adaptation

Despite the scale of the challenge, there are reasons for cautious optimism. Unlike previous technological revolutions, societies can already observe the early effects of artificial intelligence and learn from them. Countries on the front line of disruption offer valuable insights into what works and what does not.

Some governments have begun to respond proactively. National strategies focused on artificial intelligence aim to retrain workers, modernise education systems and support domestic technology sectors. Emphasis is increasingly placed on skills that machines struggle to replicate: critical thinking, complex problem-solving, communication and creative decision-making.

Education alone, however, is not sufficient. Access to digital infrastructure remains a significant barrier. Billions of people worldwide still lack reliable internet access, effectively excluding them from participation in the digital economy. Expanding broadband coverage, reducing the cost of devices and improving digital literacy are essential steps towards inclusive growth.

Investment in artificial intelligence infrastructure must be matched by investment in people. If technological capability outpaces human adaptation, inequality will deepen. Conversely, if societies can align innovation with broad-based skill development, the gains from AI may be more evenly distributed.

At a deeper level, artificial intelligence raises fundamental questions about the structure of capitalism itself. If productivity can increase without corresponding employment, how should the resulting wealth be shared? Who should govern the development and deployment of AI systems? And what obligations do those who benefit most from automation have towards those displaced by it?

Rethinking Value in an Automated Economy

The prospect of an economy in which large numbers of people have little of traditional economic value to trade is unsettling. Wages have long been the primary mechanism through which individuals participate in capitalism. If that link weakens, new forms of distribution may be required to sustain social cohesion and economic demand.

Ideas such as expanded social welfare, universal basic income and alternative ownership models are increasingly discussed in this context. While controversial, they reflect a growing recognition that existing institutions may be ill-suited to an AI-driven economy.

Ultimately, artificial intelligence is neither inherently liberating nor inherently destructive. Its impact depends on how it is developed, deployed and governed. Left to market forces alone, it is likely to reinforce existing inequalities, concentrating wealth and power among those who already possess them. With deliberate policy choices, however, it could support higher living standards, reduced drudgery and new forms of human creativity.

The central question is not whether artificial intelligence will transform capitalism—it already is—but whether societies can adapt fast enough to ensure that transformation benefits the many rather than the few. The answer will shape the future of work, inequality and economic stability for generations to come.