AI: From Employment to Jobs

Andrés Ortega

6 mins - 1 de Junio de 2023, 07:05

It has been predicted for some time. Now, even The Economist, like many others, has no hesitation in stating that the great creative models of AI (artificial intelligence) will transform lives and labour markets. This is something that governments, including Spain’s, are beginning to worry about seriously. Vice-president Nadia Calviño foresees a “profound transformation” in employment due to AI. Based on prospective studies, we have been warning about this for some time, both in terms of the automation of tasks (rather than jobs) and new ones that emerge. Automation means robotisation in a broad sense (machines and/or programmes, all with a strong AI component). Will it move towards fewer jobs, as Keynes foresaw a century ago? Be as it may, it certainly will change society and the way public goods and services are financed.
The idea and reality of employment is relatively recent. Although it has other precedents, it has accompanied the development of the industrial revolution and of commercial expansion. Compared to other realities and other methods of organising society, such as slavery (from ancient times including those of the supposedly democratic Athens), medieval serfdom, or guilds, to name but a few historical precedents. The idea of employment may be reduced or even disappear. We are moving towards a multiplicity of jobs, often performed by the same person for other people or companies – and not for everyone – or rather, the performance of tasks. In fact, the best prospective research focuses on the impact of AI and automation on tasks rather than on employment. 

[Recibe los análisis de más actualidad en tu correo electrónico o en tu teléfono a través de nuestro canal de Telegram]

To talk about these prospects at a time when employment is growing in many developed societies, when there is a lack of trained professionals for new technologies or even for old jobs, albeit with modern techniques, such as in the hotel and catering industry, may seem audacious. However, it is worth starting to think that it is possible, even probable, and that new forms of society must be devised.
A parenthesis: there is going to be huge international competition for talent, thus changing migration patterns. But the talent needed today may not be the talent needed tomorrow (e.g., coding, which is becoming more automated). The World Economic Forum’s (WEF) Openness Index forecasts that by 2040, the supply of international talent will reach an all-time high, but it will be found in new places. Destinations with weaker cultural, geographic, or diaspora ties may be left behind, the report says.
The impact of AI and automation is not mere speculation about the future. The World Economic Forum estimates that almost a quarter of all jobs (23%) globally will change over the next five years, with, in 45 economies covering 673 million workers, 69 million new jobs created, and 83 million jobs cut. That poses a net reduction of 14 million jobs, or 2% of current employment. There is more. According to the Forum’s studies, the human-machine frontier is shifting to new terrain. While expectations for the displacement of physical and manual labour by machines have diminished, tasks requiring reasoning, communication and coordination – all characteristics with a comparative advantage for humans and very middle class – are expected to become more automatable in the future.

In fact, it is already happening. There are trajectories. Carl Benedikt Frey and two colleagues at the Oxford Martin Institute have conducted an interesting study on the UK, concluding that automation in industry between 1993 and 2015 has had a greater negative impact on employment than offshoring manufacturing to China. One additional robot per thousand workers reduces employment by 0.5 percentage points. 21,109 workers in the UK were displaced in favour of China, but the introduction of robots displaced 32,328 workers. Of course, these trends need to be offset by contributions from new sectors, such as the green economy.
Beneath the good overall employment data, elements of these trends can be seen. On the one hand, the growth, albeit with ups and downs, of freelancers or the self-employed (or unpaid) – even if it is called employment – is only partly so across the OECD.
On the other hand, employment data in Spain, for example, in the last EPAS, indicate that employment has grown more than the hours worked, e.g., that there is a certain distribution of work, to the detriment of productivity. This is a phenomenon found in other countries such as Japan, which also has a declining population, although wages and productivity per hour worked have increased there. But in the OECD as a whole, the number of hours worked has fallen between 2018 and 2021 (sharply, of course, in 2020, the year of the pandemic). 
Finally, a study by the BBVA Foundation and the Valencian Institute of Economic Research (IVIE) has shown that since the beginning of the century in Spain, the value of human capital per capita has fallen by 19% and may fall by the same amount between now and 2050, due to the ageing of the population. This, among other factors, leads to greater investment in automation, e.g., in physical capital, and to a large extent, like AI, in intangibles. Although this ageing may perhaps remedy the high youth unemployment rate, which in Spain’s case is 12.8%, double the EU average.
One of the great experts in the field, David Autor, distinguishes in a recent study between innovations that increase the labour force and those that automate it. The economist sees that there is new work, new specialities introduced since the Second World War, but that in recent decades greater high-paying occupations have been generated and at the same time more services with lower wages. This, added to the jobs lost to automation/digitisation – for example, in sectors as different as automotive and banking – generates a phenomenon of a hollowing out of the middle classes, something that Trump understood well enough to get to the White House. It has not been said that the repatriation of activities promoted by the former and now by Biden will generate much more employment, as many of the tasks that return or are developed are automated. Europe does not escape these trends, which generate processes of dispossession, as Christophe Guilluy calls it in the case of France. Not forgetting that part of AI is fuelled by low-skilled and low-paid human labour – which is feeding data – which increases inequalities twofold. In other words, this exacerbates the economic, sociological, and political impact. At the bottom is the problem of who is going to pay taxes to maintain a certain welfare state and public investment in the future: the machines and the programmes, in other words, after all is said and done, capital?
Se puede leer el artículo original en español
¿Qué te ha parecido el artículo?