AI won’t take all our jobs, judging by history and economics

 In many cases, it will complement human labour rather than replace it outright.

From Carmelo Ferlito

Every major technological breakthrough brings with it a familiar wave of anxiety.

Today, artificial intelligence (AI) has taken centrestage as the supposed forerunner of mass unemployment. From automated writing tools to self-driving systems, the fear is that machines are not just replacing tasks, but entire professions.

The impression is that this time, unlike before, the jobs simply will not come back.

It is an understandable concern. But it is also one that history and sound economic reasoning suggest is deeply misguided.

The idea that technology destroys jobs faster than it creates them is not new. In the early 19th century, English textile workers known as Luddites smashed mechanised looms out of fear that automation would render their skills obsolete.

Later, the spread of industrial machinery, electricity and eventually computers sparked similar predictions of permanent joblessness. Yet in each case, while certain occupations declined or disappeared, overall employment continued to grow and living standards rose dramatically.

Why has this pattern repeated itself so consistently? A useful lens comes from the Austrian school of economics, which emphasises the dynamic, ever-evolving nature of market economies.

From this perspective, the economy is not a static system with a fixed number of jobs to be distributed. It is a process driven by human creativity, entrepreneurship and the constant discovery of new ways to satisfy human wants.

When a technology like AI increases productivity, it enables businesses to produce more with less input. At first glance, this seems to threaten jobs: if fewer people are needed for a task, some will be displaced.

But this is only part of the story. Increased productivity lowers the cost of goods and services, which in turn lowers prices or improves quality. Consumers benefit through higher real purchasing power.

And here is the crucial point: human wants are not fixed or finite. As purchasing power increases, people demand more goods and services, including entirely new ones that did not previously exist.

This creates opportunities for entrepreneurs to develop new products, industries and forms of employment. The labour “freed” by automation does not simply vanish into unemployment; rather, it becomes available for these new ventures.

Consider agriculture. In the early 1900s, a large share of the workforce in many countries was employed on farms. Today, thanks to mechanisation and improved techniques, only a small fraction of workers is needed to produce vastly more food.

If technological unemployment were a permanent phenomenon, this massive displacement should have led to widespread joblessness.

Instead, those workers moved into manufacturing, and later into services, technology, healthcare, education and countless other fields earlier generations could scarcely have imagined.

The same logic applies to AI. While it may automate certain tasks, especially routine or repetitive ones, it also opens the door to new roles that leverage human skills in different ways.

In many cases, AI will complement human labour rather than replace it outright, increasing the productivity and value of workers who can effectively use these tools.

Just as spreadsheets did not eliminate accountants but changed the nature of their work, AI is likely to reshape jobs rather than eradicate them wholesale.

The Austrian school also highlights the importance of price signals in coordinating this process. When AI reduces the need for labour in a particular sector, wages in that sector tend to fall relative to others. This sends a signal to workers and employers alike, encouraging a reallocation of labour toward areas where it is more highly valued.

At the same time, lower production costs in AI-enhanced industries reduce prices, stimulating demand and freeing up income to be spent elsewhere. The result is a continuous adjustment process in which resources, including labour, flow to their most productive uses.

None of this is to deny that technological change can be disruptive. Transitions can be difficult, especially for workers whose skills are closely tied to tasks that become automated.

The adjustment process takes time, and not everyone benefits equally or immediately. But this is a challenge of adaptation and policy, not evidence of an inevitable march toward mass unemployment.

It is also important to remember that AI itself does not act independently. It is developed, deployed and refined by entrepreneurs responding to market incentives.

If an AI application does not create value for consumers, it will not be widely adopted. Conversely, when it does create value, it tends to expand economic activity rather than contract it. New industries emerge, existing ones grow and demand for complementary goods and services increases.

The deeper mistake in the technological unemployment narrative is the assumption that there is a fixed amount of work to be done: a “lump of labour” that machines are gradually taking over.

In reality, work is defined by human wants, and those wants evolve and expand over time. As long as people continue to desire better, cheaper and more diverse goods and services, there will be opportunities for human effort to be applied productively.

AI will undoubtedly change the nature of work in profound ways. Some jobs will disappear, others will be transformed and entirely new ones will emerge.

But if history is any guide, and if we take seriously the insights of market economics, there is little reason to believe that AI will usher in an era of permanent mass unemployment.

The more likely future is not one without work, but one in which human labour continues to adapt, specialise and create value in ways we have yet to fully imagine. - FMT

Carmelo Ferlito is an economist and the CEO of the Center for Market Education.

The views expressed are those of the writer and do not necessarily reflect those of MMKtT.