The impact of Artificial Intelligence on employment: a new economic paradigm

16 February 2024

AI can be both a disruptor and an enabler of human welfare. What will be its effects on labour markets, social welfare and global value chains? The right policies will be crucial to make it economically beneficial

image credit:

By Mahdi Ghodsi

This article covers the following aspects:

  • The evolution of Artificial Intelligence (AI) and its capabilities: a historical journey
  • Two perspectives: AI as a potential threat and a tool for enhancing human welfare
  • AI's influence on 'creative destruction' and job creation
  • Advocating for a balanced and inclusive approach to AI's role in the economy
  • Technology and labour: insights from a wiiw study on wages and income shares
  • Diffusing the impact of technologies through global value chains
  • Navigating the future: the balanced potential of AI

The evolution of Artificial Intelligence (AI) and its capabilities: a historical journey

With AI, for the first time, humanity has managed to create a new intelligent entity with unique analytical capabilities. Following the invention of electricity during the Second Industrial Revolution and the subsequent development and production of complex and highly capable semiconductors (microchips), humans have progressed from analysing data in its simplest form (i.e. binary information) to designing and using the most complex analytical algorithms by the end of the Fourth Industrial Revolution. Today, with the advent of Generative Pre-trained Transformers (GPTs), it has become possible to automatically analyse complex data and questions in the shortest possible time. The new generation of AI tools is capable of processing a wide range of information and data in written, spoken and visual formats. By analysing graphical data, AI can now identify people and objects accurately and with minimal error.

Two perspectives: AI as a potential threat and a tool for enhancing human welfare

There are two dominant perspectives on the substantive philosophy of AI. The most extreme pessimistic view posits that the advancement of these tools, coupled with the significant development of quantum processing hardware, heralds the emergence of a new entity. In a rather dystopian scenario, this new entity or species, possessing informational and analytical processing power superior to that of humans, is seen as a potential threat to human dominance on Earth, possibly leading to the extinction of mankind.

In the early stages of the rise of such a superhuman entity, industrial and service robots would replace human labour in production, centralising the generated income in the hands of capital owners. In the recent book by Daron Acemoglu and Simon Johnson titled ‘Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity’, this kind of modern technology deployment is described as ‘so-so automation’, offering limited widespread benefits to human society.

The control of this technology lies in the hands of a wealthy minority composed of those who hold power over extractive political institutions in non-democratic countries and the owners of ‘superstar’ and revisionist corporations in democratic countries. By harnessing these technological tools, these minorities can easily facilitate the process of surveillance, control and repression of society according to their wishes, leading to an increase in authoritarian power by rulers or superstar corporations in these two types of countries. In democratic countries, however, the control of superstar corporations can be regulated by governments elected by a majority.

If democratic and popular control over these powerful technological tools is not maintained, there is an increased likelihood that AI will prevail over humans, as AI has not been shackled under the control of societal powers since its inception. This scenario has been depicted in many comic stories since the advent of computers in the Third Industrial Revolution.

On the other hand, from an extremely optimistic perspective, by entrusting the power of inclusive institutions to society and consistently formulating regulations to govern the market for the production and consumption of these technologies, humanity can make optimal use of them to enhance human welfare. Acemoglu and Johnson critically discuss this optimistic view, referred to as the ‘productivity bandwagon’, suggesting that technological advancement leads to increased overall productivity and higher wages in the lower deciles of the income distribution. Despite the disruptive role of new technologies in replacing many traditional jobs, this perspective maintains that these technologies can play a positive and increasingly significant role in enhancing societal welfare through two distinct channels.

AI's influence on 'creative destruction' and job creation

First, these technologies lead to the phenomenon of ‘creative destruction’, a term first coined by the Austrian political economist Joseph Schumpeter, who was inspired by the theories of Karl Marx. According to this phenomenon, the introduction of new production methods displaces the old ones. This process has triggered popular resistance in all industrial revolutions. The Luddite movement in the 19th century, involving textile workers who destroyed new textile machinery in England, is a famous example. As another example, during the First Industrial Revolution (1760-1840) in Britain, when steam power started to replace horse-drawn transports in 1825, the Austro-Hungarian Empire refrained from developing railways, fearing that horse and carriage owners would revolt. However, the increased productivity in railway transportation in the British Empire led to significant growth in the production and export of agricultural products, textiles and industrial goods. This, in turn, resulted in faster development and expansion of national income and aggregate demand, ultimately leading to broader employment compared to the Austro-Hungarian Empire. Thus, the first positive channel through which new and dynamic technologies can contribute to societal welfare is the increase in aggregate demand resulting from enhanced productivity.

Indeed, today’s cutting-edge AI technologies are capable of replacing a large portion of simple and routine jobs (e.g. customer service responses as well as order and data entry), along with certain aspects of assembly and production lines in various industries. However, in addition to being able to replace many jobs, AI has also created new job roles and tasks. At the 2023 World Economic Forum, Richard Baldwin made a noteworthy point, saying: AI won’t take your job; it’s somebody using AI that will take your job.

At the same time, it’s important to remember that all jobs related to information and communication technology (ICT) only emerged in the decades following the Second World War with the advent of the Third Industrial Revolution. In 2021, employment in the ICT sector was equivalent to 3.7% of total employment in OECD countries. Therefore, as a second positive channel, AI technology can play a more effective role in enhancing welfare, not only by creating new jobs and tasks with higher incomes compared to other jobs but also by increasing productivity in other industries and boosting overall demand.

Advocating for a balanced and inclusive approach to AI's role in the economy

The truth does not lie entirely in either the extremely pessimistic or the extremely optimistic viewpoint. Despite the wide variety of governance systems in various societies and countries, technological progress moves along a path between these two extreme perspectives. Countries that have the ability to harness, produce and utilise AI technology in a way that fully serves and is controlled by society will be better able to reap the full range of benefits of these technological advancements. Indeed, the transformative power of technology has long been the cornerstone of economic evolution, reshaping industries and labour markets. Now, with automation and digitalisation occurring across various industries, important questions are arising about their impact on wages and the labour force’s share in the total added value.

Technology and labour: insights from a wiiw study on wages and income share

In our recent wiiw study titled ‘Assessing the Impact of New Technologies on Wages and the Labour Income Share,’ we address questions regarding the globalisation trend of the economy as well as the nuanced and dynamic differences in the labour markets of 32 developed countries, considering all 38 industries within them. The study focuses on examining the effects of technology spill-over in global value chains on the labour income share of total value added and wages. Econometric findings of the study indicate that innovation, as measured by patent grants, correlates positively with wage increases. This supports the hypothesis that innovation can bring substantial rewards for the workforce – especially for those members actively engaged in innovative fields.

However, the protection of intellectual property rights and patent laws by capital owners enhances the quality and supply of innovative products. This, in turn, increases sales, market share and, ultimately, profits and innovation rents in that industry. Admittedly, this issue could lead to an increase in capital’s share of the added value relative to labour’s share. Nevertheless, the negative impact of this issue on the labour income share of added value in an industry can be offset by the positive effects of innovations provided by international suppliers in the same industry.

Diffusing the impact of technologies through global value chains

At the same time, the adoption of industrial robots in an industry’s production process can increase the labour income share of its total income. This phenomenon illustrates the operation of the ‘productivity bandwagon’ theory in these countries. Additionally, deploying more robots in the international industries that supply it can also boost the labour income shares of an industry’s total income. Conversely, the impact of employing industrial robots in supplying industries, as well as in domestic buyer sectors and among international buyers of an industry, is negative on the labour income share of that industry.

Despite these opposing forces, the aggregate effects resulting from the utilisation of industrial robots across all industries via existing global channels support the theory of ‘so-so automation’. This theory posits that the labour income share of the total income in these industries has experienced a noticeable decrease in these countries.

In most models examined in this analysis, investment in ICT assets (e.g. computers and digital infrastructures) has not been shown to have a significant impact on wages and salaries. However, as the share of digital capital in total capital increases through the process of digitising capital, overall demand has risen. This, in turn, has heightened the demand for labour. Ultimately, the increase in demand has boosted the aggregate income of the workforce and, consequently, the level of welfare in those countries, thereby supporting the ‘productivity bandwagon’ theory.

Navigating the future: the balanced potential of AI

As a conclusion, it can be said that the true potential of AI lies in a balanced approach, avoiding both excessive pessimism and blind optimism. This article emphasised the importance of inclusive institutions and strategic regulations in harnessing AI’s power for societal benefit. The review of literature and research findings suggests that while AI may disrupt traditional jobs, it also creates new opportunities and has the potential to increase overall productivity and welfare. As Acemoglu and Johnson (2023) note, the key lies in navigating technological evolution thoughtfully, ensuring that AI advancements contribute positively to the economy and society.