Have Skills, Will Travel

As a new CESifo study by Eugenia González Ehlinger and Fabian Stephany shows, employers have started focusing on individual skills rather than on formal qualifications in their recruiting. To find this out, they trawled through a large dataset of around one million UK online job vacancies between 2019 and 2022, and perused a large body of diverse literature on technological change and labour market signalling.

Their results provide evidence that employers have begun to resort to so-called “skill-based hiring” for AI and green roles, as more flexible hiring practices will allow them to increase the available talent pool. Over the period of their study, the demand for AI roles grew twice as much as average labour demand, with AI roles advertising five times as many skills as job postings on average.

What’s more, the mention of university education as being required for AI roles declined by 23%. In other words, university degrees no longer offer an educational premium for AI roles. Interestingly, however, for green positions the educational premium does persist.

Our researchers recommend making use of alternative skill building formats, such as apprenticeships, on-the-job training, massive open online courses (MOOCs), vocational education and training, micro-certificates, and online bootcamps to use human capital to its full potential and to tackle talent shortages.

After all, now that skills appear to trump formal education, AI skills have been found to offer a wage premium of 16%, just a bare whisker below a PhD.

Another CESifo paper, by Daniel Goller, Christian Gschwendt and Stefan C. Wollter, looks at the effect of artificial intelligence in labour from a different perspective. Their findings show that the deployment of generative AI, in the form of ChatGPT, has triggered a strong and long-lasting decline in the intensity of searches for apprenticeship vacancies, suggesting a great deal of uncertainty among the affected cohort.

When young people get the apprenticeship of their choice, they invest a large part of the next three to four years learning occupation-specific skills that will enable them to perform specialised tasks. But they now seem to worry that, if technology can easily perform these tasks now or in the near future, this investment will likely be lost.

In the past, digitalisation was skill-biased, shifting demand from less-skilled to more-skilled workers and leading to a skill upgrading of the labour force. Furthermore, digital technology increasingly substituted workers in routine-intensive jobs, while complementing workers executing non-routine tasks. As a result, many labour markets experienced a decline in routine employment and a steady growth in non-routine cognitive jobs, which predominantly require higher levels of education. Occupations became more demanding and complex and thus required longer, higher, and better training.

The use of generative AI, however, leads to increased productivity in more complex and less ambiguous tasks. Early estimates suggest that, of the many occupations that will be affected by AI in the near future, it is those currently performed by higher-educated workers that will be disproportionally exposed, particularly in tasks demanding literacy/reading, since almost 60% of the workforce using literacy skills daily have a proficiency comparable to or below that of computers.

Worryingly, other studies show that the demand for freelancers providing work that AI tools are specialised in significantly decreased, while their employment and earnings suffered. The negative impact is greater on freelancers who did higher-quality work before the tools were introduced.

Small wonder then that the cohorts that will be tomorrow’s workforce are feeling insecure regarding their occupational choices. With the launch of ChatGPT, the supply of potential apprentices, as represented by search queries for apprenticeship positions, decreased substantially, by about 8 percent on average.

Thus, a slight conundrum: on the one hand, acquiring such specialised skills as those for AI appears to be the safer bet for higher wages, on a par with getting a phD. On the other, potential apprentices appear unsure which path to pursue, keeping them from starting the acquisition of such skills in the first place.

An ideal setup, if there ever was one, for getting policymakers, educators and employers around the same table to chart together a path forward.

Maybe they could invite ChatGPT as well.