AI creating not killing jobs – European Central Bank

The adoption of the technology is correlated with reduction in wages, according to newly published research

The widespread adoption of artificial intelligence and related technologies has led to an increase in human jobs, but a decrease in wages, according to a research bulletin published by the European Central Bank on Tuesday.

Titled “Reports of AI ending human labour may be greatly exaggerated,” the paper purports to counter fears of AI decimating the labor market with data from 16 European countries taken between 2011 and 2019. 

“During the deep learning boom of the 2010s, occupations potentially more exposed to AI-enabled technologies actually increased their employment share in Europe,” the paper states, acknowledging that “the jury is still out” on whether that pattern will hold true in the future. 

Read more

The employment share of economic sectors most exposed to AI increased, according to the ECB’s research, with high-skilled positions – and especially those held by younger workers – seeing the biggest boost. AI exposure was twice as likely to benefit workers in the youngest third of the population, the research showed.

However, at least one study cited in the paper showed that individual employers within those AI-exposed sectors reduced hiring for non-AI-related positions – and hiring in general. The effect on low- and medium-skilled jobs was less pronounced, with no indication that software was replacing routine-heavy positions, though earlier studies mentioned by the writers suggested many of these jobs may have been lost in previous years to less advanced forms of automation.

While two of the three studies examined in the paper showed no meaningful relationship between AI exposure and wages, the third found “neutral to slightly negative impacts” on human earnings, revealing that occupations most exposed to AI showed worse wage growth than those insulated from the technology. 

The writers acknowledge that the degree of labor market disruption by AI varied significantly between countries, with some – the paper does not name which – affected adversely by AI-enabled automation, in contrast with the larger trend. 

The ongoing and unpredictable development and adoption of AI and related technologies means “most of their impact on employment and wages – and therefore on growth and equality – has yet to be seen,” the researchers wrote.

The ECB has its own reasons for projecting a rosy future for AI, having announced in a September blog post that it would explore the use of the technology in economic modeling and data crunching, informing everything from routine economic analyses to critical decision-making processes.

The bank’s chief services officer Myriam Moufakkir pledged at the time to “accelerate” the adoption of AI across all applications to keep the ECB “modern and innovative” while safeguarding the privacy and other legal rights of all involved entities.