The automation of work through the use of technology historically introduced fears about the effect these processes would have on employment. In particular, since the industrial revolution, debates about the impact of technological advances on the production of goods, to the detriment of traditional labour, have intensified to the point of becoming movements. If we look back 200 years, we already find antagonistic tendencies such as Luddism and machinism that confronted each other in their prediction about the future of work from the introduction of machines in production processes.

However, despite the incorporation in factories of packaging lines, heavy machinery or computers that replace the physical and calculating work of individuals, the number of people who managed to enter the labour market has been growing for 200 years (with some volatility) and the jobs destroyed have eventually been reallocated to other sectors. We could describe this phenomenon as a process of "Creative Destruction".

If we turn the calendar forward to the 2010s, we see an automation drive emerging that raises new warnings about high risks of technological unemployment. This time, the technologies that would come to automate work are artificial intelligence and high-precision robotics. In 2013 Frey and Osborne, two researchers from Oxford University analysed more than 700 occupations and concluded that in developed countries, 47% of occupations could be automated within a couple of decades (go to paper).

In the years that followed, the debate on these prospects became broader and deeper, adding important nuances. It is also necessary to bear in mind that narratives are developed that meet public policy needs and that these differ from region to region.

On the one hand, in discussions following Frey and Osborne's work, it is becoming clear that few occupations would be fully automatable, since in reality it is a matter of substituting tasks within the occupation. This naturally transforms the occupation and requires workers who are able to integrate with the digital tools of automation, which is where one of the central concepts for vocational training, upskilling, comes in. In order for the workers themselves to be able to perform in the job, they require training, but it is training at a higher qualification level, hence the term "upskilling". Those workers whose occupations are radically automated will require reskilling, in this case training (now) in other occupations.

In some sectors and for some occupations, most work requires specific skills. Simply automating some skills does not imply that all the skills deployed in the job become unnecessary. In fact, one might think the opposite: by not having to devote time and effort to automatable tasks, other activities and skills become relevant, thus enhancing the value contribution to the production process. This is apparently true in high-skilled occupations, but less true as we move towards the lower skill levels.The automation of routine tasks automatically places greater value on the expertise, creativity and human (socio-emotional and transversal) competences of workers, compared to the recent past; so as automation frees up our time, we can tackle more complex problems, new products and services that stimulate use and consumption; and hence demand. There is an expectation that this demand will also be met by the creation of new jobs.

On the other hand, Frey and Osborne make it clear that the automation risk they refer to is analysed from the angle of available technologies. Simply put, they indicate that in a couple of decades the technology will exist to automate those occupations, which we now know to be tasks. This caveat is important, because the processes of work automation depend on a number of non-technological variables such as the speed of technology adoption in an economy, social and tripartite agreements.

Indeed, there are studies that show some correlation between the introduction of robots and falls in employment-to-population ratios (in routine and "codable" positions), and wages in the US. In the worst-case scenario, where the stock of robots quadruples by 2025, we would see 1% reductions in the employment-to-population ratio and lower wage growth of between 1.3 and 2.6%.

According to VOXEU, these effects are considerable, but they would impact a relatively small fraction of the American workforce, so it is possible to say that there is not yet consistent evidence to prove that in the future new technologies will make most jobs disappear

Another study on the incorporation of robots in Germany does not detect an impact on employment at the aggregate level. The study argues that the role of business associations and trade unions appears to be key to making this process less socially disruptive. The same study finds evidence of falling job quality and greater income inequalities within productive sectors.

In short, the process of labour automation has not stopped for at least 200 years. The current risk of automation, based on artificial intelligence and high-precision robots, threatens the replacement of cognitive skills as never before, which is why there is a certain level of alarm greater than in previous technological revolutions.

What we do know is that where automation is occurring, it is displacing workers who are engaged in more routine and therefore lower-skilled/lower-paid tasks. The cases of Germany and the US show that there are greater inequalities and losses in job quality. Perhaps the most important point to highlight in this phenomenon has to do with the increase in inequality as well as the risk of unemployment.

From the perspective of vocational training, the mandate is clear: upskill, reskill and in particular train all new generations in skills to move in a changing labour market; where movements within one's own sector may already be very challenging, but where the likelihood of having to move not from one job to another but from one sector to another is increasing.