How to stop AI from becoming the enemy of younger workers

Viewed from afar, South Korea looks like a clear winner from the rise of artificial intelligence. Some of its biggest companies are booming because of the global appetite for chips and data centres. Samsung Electronics and SK Hynix, its leading chipmakers, have both topped US$1 trillion (S$1.29 trillion) in market capitalisation.

And because these companies have powerful unions, their workers are securing a share of the surge in profits. For example, in May, Samsung and its union came to a profit-sharing agreement that is expected to award employees in the memory chip division an average bonus of nearly US$400,000. Little wonder that South Korea is one of the few large countries where optimists about AI outnumber pessimists, according to polling by the Pew Research Center.

But there is more to the story. South Korea has long been an economy with deep fractures: between “insiders” with secure jobs in large, productive companies and “outsiders” with insecure jobs in small and medium-sized enterprises (SMEs). So far, the benefits of AI seem to be following and even deepening these fault lines. The Samsung profit-sharing agreement, for example, does not benefit the workers of the company’s subcontractors, who typically receive lower pay and less security.

Then there are those “outsiders” who have yet to get a foothold in the labour market. A study of administrative data by economists Jinsu Han and Samil Oh of the Bank of Korea found that 211,000 youth jobs (for those aged 15 to 29) disappeared over the past three years, while employment for workers in their 50s increased by 209,000 over the same period.

Youth employment fell particularly sharply in AI-exposed sectors: by 11.2 per cent in computer programming and system integration and management, 20.4 per cent in publishing, 8.8 per cent in professional services and 23.8 per cent in information services.

The economists concluded that the labour market was “undergoing a seniority-biased technological change” because AI “more readily replaces tasks performed by junior workers that rely on codified, textbook knowledge, while it tends to augment tasks requiring career-based tacit knowledge and social or interpersonal skills, which are more common among senior workers”.

There have been hints of this “seniority-biased” pattern emerging in the US too, especially in the software profession. But the trends in South Korea look particularly stark. Is this a warning for other economies? I think so.

But South Korea’s “insider-outsider” labour market structures have probably played a part too. Jiyeun Chang, senior research fellow at the Korea Labor Institute, says that “incumbent workers in large firms enjoy strong employment protection, so when firms adjust to new technology, the adjustment falls disproportionately on the hiring margin – that is, on young people trying to enter”.

Indeed, young people in South Korea were struggling in the labour market even before ChatGPT’s release in late 2022.

A report by the Organisation for Economic Cooperation and Development that year found that the large companies that once drove job creation were shifting towards “a more capital and technology-intensive product mix” and relying more on outsourcing, but that graduates who were unsuccessful in their attempts to get hired by the big companies or the public sector tended to “queue for such jobs rather than fill labour shortages at SMEs”. SMEs, meanwhile, were “trapped in a vicious circle”: not productive enough to offer good pay, and therefore unable to attract high-quality workers who could boost productivity.

The good news for South Korea is that the government has money to spare, thanks to rising tax revenues from the semiconductor boom. It is forming plans for a “Future Response Fund” to invest in mega-projects, address inequality and provide employment support for people in their 20s and 30s.

One obvious option would be to use some of the windfall money to subsidise companies to hire and train young people, even if their labour cannot easily be monetised until they have acquired more expertise.

A more imaginative policy agenda could extend security, training and capital to “outsiders” such as the self-employed. In the right conditions, couldn’t South Korea’s tech-savvy youth create companies of their own that use AI to do new things? Start-ups that could one day even disrupt those corporate giants that denied them a foot in the door?

New eras require new mantras. For young people who are struggling to get hired the traditional way, this might end up being theirs: If you can’t join them, beat them. FINANCIAL TIMES