Fears of job losses and large-scale unemployment increases as a consequence of the use of artificial intelligence (AI) applications by businesses have so far not been borne out, as studies conducted in Europe and the US show.
Labour market changes are occurring and are expected to continue to intensify as the diffusion of new AI applications is gradual, starting with larger firms that have the resources for the necessary investments. However, overall employment is likely to increase rather than decrease as a result of these changes,
On the one hand, the demand for occupations that can be substituted by creative AI models is decreasing, but on the other hand, the demand for new occupations that are relevant to it is increasing.
According to research by the World Economic Forum, AI is projected to create 69 million new jobs by 2028. These require skills in areas such as data analysis, machine learning and new software development, highlighting the importance of education – through university studies and retraining programmes – to obtain them.
The most important reason cited by those who predict that overall employment will increase with AI is the productivity gains it brings, which are already being felt and are expected to increase further in the future. The higher the productivity, the higher real wages will be and the stronger the economic activity, with a positive impact on the level of employment. This has been true in times of major technological change in the past and can reasonably be assumed to be true in the future. McKinsey, for example, estimates that AI could contribute up to $13 trillion to the global economy by 2030.
According to a survey of 12,000 European firms by the Center for Economic Policy Research (CERP), AI has increased productivity by an average of 4%, but with wide variations in the distribution of benefits. The biggest gains have been made by large and medium-sized companies and those that have been able to integrate AI into their production process and human capital.
Morgan Stanley also expects AI to increase productivity and real wages. It finds that unemployment has risen modestly in categories of workers most exposed to AI, Unemployment among young people aged 22-27, who are more likely to do automated jobs, has seen the largest increase since 2023 in occupations most vulnerable to AI. But if this age group is excluded, there is little evidence of widespread labour market disruption. There is, however, a source of concern from the fact that firms, when presenting their financial results, are increasingly reporting job losses due to AI rather than job creation.
Back in the past, economists at Morgan Stanley looked at five major waves of innovation in the US – from the industrial revolution to the internet – to measure their impact on the economy and the labour market, finding that technological change has always worked as a booster rather than a detractor from overall employment.
During the industrial revolution, the productivity of the US economy grew by about 0.84% per year from 1800 to 1850. Employment in the agricultural sector fell from 75% to just over 50%, but jobs in manufacturing and construction more than doubled between 1820 and 1850.
In the era of steam, rail and steel (1830-1910), often referred to as the second industrial revolution, productivity growth approached 2% a year by the end of the 19th century, twice as much as in the first industrial revolution. Employment in the agricultural sector continued to decline, falling to 30% in 1910, but employment in manufacturing increased dramatically.
In the electric and automobile revolution (1890-1950), productivity in the US economy as a whole grew at an average rate of 1.5% from 1909 to 1929, but productivity outside the agricultural sector doubled over the same period. Employment in the US agricultural sector had declined to about 20% by 2040, and from 1910 to 1950 civil service jobs nearly tripled.
In the era of electronics and aviation (1940-1980), productivity was growing 2.5%-3% per year and the service sector became dominant.
The service sector was dominant.
Finally, in the era of the internet and digital networks, productivity accelerated to 3% per year from 1990 to 2000, while demand for workers for software, data science and cybersecurity increased.
