[Heo Manseop's Deep Learning] 2.55 Million 'Resting'... Are Job Policies Also Idle in the Age of AI?

From Coding and Reasoning to Physical AI

"AI to Take On All-Around Tasks Starting in 2030"

Youth Employment Issues Overshadowed by Real Estate

Need for Solutions Such as Reduced Working Hours

Recently emerging artificial intelligence (AI) systems have gone beyond generating text, images, and video, extending their capabilities to coding, analysis, and reasoning. AI now supports complex decision-making tasks even in specialized professions such as law and accounting. 'Physical AI,' which acquires a body to move and perform tasks directly, is driving automation and productivity innovation in factories. The stock prices of SK hynix and Samsung Electronics, which manufacture memory semiconductors for AI, are surging. Taken together, these developments suggest that we may already have entered the era of the Fourth Industrial Revolution.

Converting the 1932 artwork "Lunch atop a Skyscraper" into an AI-generated image

The core of modern society is 'intelligence,' with AI inevitably at its center. AI's influence is especially significant in the economic sphere. While many focus on the benefits AI brings, it is not without side effects. Chief among the anticipated problems is a reduction in jobs. There is considerable likelihood that AI-based automation will decrease the number of available jobs. In fact, this is already becoming a reality. At the 2016 Davos Forum, it was predicted that the AI revolution would cause 7.1 million jobs to disappear in developed countries, while 2 million innovative new jobs would be created, ultimately resulting in a net loss of 5.1 million jobs. This forecast has since served as a key basis for concerns over a sharp decline in jobs due to AI. An analysis of 226,189 related articles from 54 domestic media outlets over the past year showed a 'strong' correlation (0.88 out of a maximum of 1.0) between AI and jobs. This suggests that linking AI to employment issues has become a standard reporting practice.

Both engineers and economists identify the potential for mass unemployment as one of the most significant social issues AI could cause. They warn that AI may create an industrial ecosystem that no longer requires humans, structurally eroding jobs. What would happen if this truly comes to pass? The 'Life Resource Management Institute' predicts the emergence of factories without workers, restaurants without chefs, sports games without referees, offices without office workers, newspapers without journalists, hospitals without doctors, and courtrooms without legal professionals. Already, in many fast-food outlets, kiosk machines are replacing human staff by taking orders.

Of course, there are strong counterarguments that AI will not necessarily reduce jobs. Hans Moravec, a Canadian researcher in AI and robotics, proposed the paradox known as 'Moravec's Paradox,' which states that "tasks easy for humans are difficult for robots, and vice versa." According to this, cooperation between humans and AI is expected to improve, and jobs will be maintained. The 'law of the invariance of total labor' also asserts that "as many new jobs are created as are lost to the industrial revolution."

However, it is uncertain whether this paradox and law will hold up in an age of intelligence dominated by extremely efficient AI. According to the research team led by Eom Hyojin, weak AI will replace humans in repetitive tasks, while strong AI will take over even proactive and high-level areas of work. Now is the time to view AI-driven job losses as a serious issue and to prepare accordingly. In fact, the crisis may have already begun. According to the Ministry of Data and Statistics, in 2024, the number of people counted as 'resting' among the economically inactive population reached 2,555,000, the highest since statistics began in 2003.

Youth unemployment, in particular, is a serious problem. The number of 'resting youths'—those in their mid-to-late 20s who are not seeking employment—has increased 2.6 times over the past 20 years (Korea Employers Federation, April 20 report). In February, the number of employed people aged 25–29 was 2,346,000, the lowest in nine years. The main causes are cited as the increase in hiring of experienced workers, a decrease in hiring in professional service industries, the mandatory retirement age of 60, a decline in job creation capacity due to entrenched low growth, and young people’s preference for large corporations. Among these, the increase in hiring experienced workers and the decrease in hiring in professional service sectors are highly likely to have been influenced by AI.

A significant number of mid-level employees or professionals with 5–10 years of experience, who are familiar with their company’s work, are able to leverage AI to handle the workload of several new hires. The productivity gap between these employees and 'job seekers with no work experience' is widening further. Moreover, AI is showing signs of reorganizing even the roles of these middle managers. The 'McKinsey Report' projects that "from 2030, AI will intervene comprehensively in corporate product development, production, marketing, finance, legal affairs, human resources, research and development (R&D), and software work." With jobs already shrinking now, there is growing concern about what will happen if this 'singularity' arrives in 2030.

Adults with a college education who cannot find employment typically face severe hardship. Even if they are willing to work, living without a job is a significant source of pain. Many job seekers overcome the fierce competition of college entrance exams and spend their entire university years accumulating language scores, grades, and credentials, only to repeatedly fail at the threshold of employment. Once the optimal hiring period passes, even the already narrow path to a job closes entirely.

The government has made efforts to increase employment. However, it is questionable whether employment issues are as prominent a national agenda as real estate issues. President Lee Jaemyung has actively used social networking services (SNS) such as X (formerly Twitter) to heat up social debates, often highlighting key social issues. Recently, he gave significant attention to posts about the abolition of the long-term holding special deduction for real estate transfer taxes. In contrast, over the past year, only one-tenth as many news reports on President Lee’s X posts focused on employment compared to those on real estate (BigKinds analysis).

Primarily, the responsibility for finding a job rests with the job seeker. Even if one cannot develop AI or software, one must at least be able to apply such tools to their field of expertise. If large corporations prefer experienced hires, it may be worth considering gaining experience step by step, rather than exclusively aiming for big companies. However, it is also difficult to ignore that the job shortage is so structural that it cannot be left entirely to individuals and that the responsibility of the state is significant.

President Lee has stated that "it is common sense that, given the same conditions, the wages for irregular workers with unstable employment should be higher." This should serve as an opportunity to reform the rigid labor market structure that blocks youth employment. The impact of the Yellow Envelope Act (amendments to Articles 2 and 3 of the Labor Union and Labor Relations Adjustment Act) and other such legislation on companies' job creation capacity should also be examined. AI policy, regardless of whether it is called 'sovereign AI' or something else, must include solutions to youth unemployment. Compared to the shock that AI transformation may bring to youth employment, current policies still lack tailored responses.

It is also time to begin a social discussion on how to distribute the increased profits resulting from AI-driven efficiency. The Samsung Electronics labor union's demand for bonuses on a per-person basis could help foster such debate. One might also argue, "If AI and semiconductors dominate the socioeconomic landscape, should the resulting gains go only to the companies and unions involved?"

In the future, one employee might do the work previously handled by five, with AI taking over the rest. In harsh terms, there are two options. The first is to lay off the remaining four employees, resulting in an 'AI dystopia.' The second is to keep all five employees while reducing working hours—a scenario of 'AI utopia.' The outcome depends on the actions of citizens, businesses, and the government. Europe's discussion of 'Responsible AI' is based on the premise that "AI should not be harmful to people." The productivity benefits of AI should not be concentrated in the hands of a few, nor should the majority seeking jobs be left frustrated.

Professor Heo Manseop, Department of Digital Media and Communication, Kangwon National University

This content was produced with the assistance of AI translation services.