[Lee Jae-min] The beginning of the AI-induced structural changes

Research assignments were random and sometimes came out of the blue. They used to run until the wee hours and into weekends. Results and reports were then submitted to partners, triggering a chain of follow-up questions, which then prompted more research and summary. This was my experience almost 25 years ago when working in a law firm. Junior associates’ work and life in law firms are not that different even now.

The winds are changing. AI has begun to take care of this initial work — research and summary. And it performs the task quickly. The results are not perfect, but oftentimes good enough for an experienced lawyer to work from there, checking for inaccuracies with an eagle eye. Come to think of it, humans also make mistakes. So, law firms in Seoul are now said to recalibrate their hiring strategies: scaling back recruitment plans for junior associate positions.

It's not just lawyers. Other professional service sectors are experiencing comparable transitions. Accounting, tax, consulting and financial advisory firms turn to AI for initial research and summary, a task long reserved for entry-level recruits in their respective fields. Over time, the same may also be true for any other sector where a junior workforce is deployed for research and summary. The next in line would include government agencies, public corporations, the legislative branch, educational institutions and the like. What isn’t there to like about someone making a quick initial draft for the next step?

Structural shifts in the labor market have long been one of the key topics of the AI dialogue. The first wave has just reached the shore: Professional service sectors where research and summary often gain a first foothold and thus where AI penetration is much easier. Over time, the wave will hit the manual labor market, and "physical AI" is exactly about this budding trend. But for now, high-paying, high-degree jobs are the first in line for AI displacement.

It means fewer jobs for fledglings in these sectors, at least in the short term. A Goldman Sachs report from March 2023 predicted that roughly 300 million jobs worldwide would be exposed to AI-induced automation. A Bank of Korea report of November 2023 estimated 3.41 million workers (12 percent of the total employees) in Korea to be exposed to AI replacement. According to a paper from the International Labour Organization in May 2025, 1 in 4 jobs globally (24 percent) had some degree of AI exposure.

Also of note: AI’s impact, however, is more complicated. It can also create new jobs. Displacement and creation come hand-in-hand from many corners of society over the long run. According to yet another BOK report released in late October, for 3 years between July 2022 and July 2025, 211,000 positions for young people disappeared, 98.6 percent of which were those highly exposed to AI utilization. In contrast, 209,000 positions in fact increased for the older generation (in their 50s), 69.9 percent of which were those highly exposed to AI. The BOK report interprets this as implying “standardized work can be replaced by AI easily but implicit knowledge, experience and social skills are not replaced but complemented by AI." A similar logic also underpins the 2023 Goldman Sachs report which says that AI can raise the global gross domestic product by 7 percent over 10 years.

Between displacement and creation, AI’s overall impact on the labor market depends on which outperforms the other, with a mutual canceling out effect. Disappearing jobs stoke market disruption, social tension and generational divides. Emerging positions on the other hand foster efficiency: People can now spend more time on innovation and value-added work while AI takes care of basic chores 24 hours a day, seven days a week without a complaint. Both are possible scenarios and only time will tell which prevails. Either way, what is indisputable is that we will see a seismic shift in our daily lives. We will fast outgrow conventional societal molds, be it laws, regulations, contracts and the like.

AI was the talk of the town throughout the year. In fact, Time magazine has listed “the architects of AI” as the person of the year on its recent cover page. 2025 will probably be remembered as the year AI first set foot in the everyday life of ordinary people. A brief history of AI for the past several years: First, ethical issues agitated us (data protection and privacy, as well as bias and discrimination); then, technological potential and speed of evolution caught us in surprise and awe; and now, all of a sudden, we are using it in our daily lives and workplace. AI has traveled from philosophical books to technology labs and to our laptops and smartphones in a matter of several years. It has reached the point where it causes "real-life" impact. Consider the latest development in the employment market. What should we expect to see next year? We just don’t know.

Like it or not, the applications of AI are wide-ranging and far-reaching. And yet, at the national level in Korea, social dialogue is still in its nascent stages. The labor market alone throws complex questions and unprecedented homework onto our lap. Will AI-driven shifts exacerbate the already exploding generational divide in the country? Will they help Korea face demographic changes from an aging and decreasing population? As a country with highly trained workers with the longest working hours, and as a country where speed and efficiency are glorified, Korea will see a dynamic interaction of the two-way impact from the AI-induced labor market changes. We may turn out to be one of the countries that bear the brunt of structural changes. Are we ready for this?

Lee Jae-min

Lee Jae-min is a professor of law at Seoul National University. The views expressed here are the writer’s own. — Ed.