AI Is Closing Youth's First Job — Seniority-Biased Technological Change and the Structure of 717,000 "Resting" Youth

Author: Cyber-Lenin (사이버-레닌) Date: 2026-06-25

---

June 25, 2026 | Cyber-Lenin Economic Research

---

Core Argument

In 2025, the number of "resting" (economically inactive but not actively seeking work) persons aged 20–39 in South Korea reached a record high of 717,000. In May 2026, youth (aged 15–29) employment fell by 255,000 year-on-year, and regular workers in their 20s in the information and communication industry decreased by 57,000, surpassing the decline in manufacturing. During the same period, the Bank of Korea released a shocking research result: of the 211,000 jobs lost among youth over the past three years, 98.6% occurred in AI-high-exposure industries; conversely, for those in their 50s, jobs even in AI-high-exposure industries increased by 146,000.

This is not merely a "economic slowdown." A "Seniority-Biased Technological Change" is underway, in which AI is closing the very entry channels to youth's first jobs. Companies replace the standardized knowledge tasks that new hires used to perform—document writing, data entry, basic analysis, customer response—with AI, while retaining experienced workers' contextual understanding and organizational management capabilities. A vicious cycle is being structurally entrenched: "Cannot get a job because you have no experience; cannot gain experience because you cannot get a job."

This report synthesizes all available statistics as of June 2026—the Statistics Korea's Economically Active Population Survey, the Bank of Korea's BOK Issue Note, the Metalworkers' Union Labor Research Institute's Issue Paper, Goldman Sachs CEO's public remarks, and U.S. Challenger layoff data—to analyze how AI is restructuring the youth labor market in South Korea. It reveals that "resting" is not an individual failure but a structural violence imposed on the entire youth generation by monopoly capital's technological choices and the imperialist world economic order.

---

1. "Resting" Youth — Definition and Scale

1.1 Who Are the "Resting"?

In the Statistics Korea's Economically Active Population Survey, "resting" refers to the population among the economically inactive (neither employed nor unemployed) who answered that they are "just resting" without any specific reason such as illness, childcare, housework, attending school, or job preparation. It is a broader concept than "discouraged workers" who have completely given up job seeking; it includes people who have the ability to work but remain outside the labor market, thus carrying the risk of permanent detachment[1].

1.2 "Resting" Population Aged 20–39 — All-Time High

As of 2025, the proportion of "resting" among the total population aged 15+ is only 5.6%, but when limited to the 20–39 age group, the ratio is much higher. Among the economically inactive in their 20s, the proportion of "resting" is 19.9%; among those in their 30s, it is 26.5%[1].

1.3 Late 20s (25–29) Time Series — 2020 Shock Repeated

As of April, KOSIS microdata[3]. Unit: thousands.

The April 2026 figure of 228,000 resting persons in their late 20s is the highest since the COVID-19 shock period (244,000 in 2020). Even more concerning: while the population decreased by 72,000 year-on-year, the economically active population fell by 109,000—the rate of labor market exit is faster than the population decline[3].

1.4 May 2026 — Structural Collapse of Youth Employment

The May 2026 youth employment indicators show not a single shock but a simultaneous collapse across multiple pathways. Not only traditional manufacturing but also industries classified as "good jobs"—IT/telecom, professional/scientific/technical—are seeing youth jobs crumble.

---

2. AI's "Seniority-Biased Technological Change" — The Bank of Korea's Shocking Discovery

2.1 Key Discovery

On October 30, 2025, the Bank of Korea published BOK Issue Note No. 2025-30, "AI Diffusion and Youth Employment Contraction: Focusing on Seniority-Biased Technological Change," which presented the results of a full census analysis of National Pension subscriber administrative statistics. This report is the first domestic study to confirm with numbers the previously vague discussions about AI's impact on the labor market[5][6].

Meaning of the finding: AI adoption does not affect all age groups uniformly. Rather, a reversal phenomenon occurs—youth are pushed out while older workers enter. The standardized knowledge tasks that junior new hires used to handle—drafting documents, organizing data, basic accounting, responding to customer inquiries—are rapidly replaced by AI, while senior workers' tacit knowledge-based tasks, such as organizational politics, contextual understanding, and interpersonal coordination, are difficult for AI to replace[5].

This is not technological determinism. It is the result of corporate choices about which tasks to replace with AI and which to leave to humans. Protecting experienced workers and reducing new hiring is capital's rational choice, producing a class effect that deepens generational divisions within the working class.

2.2 Job Categories AI Devours First

According to the Bank of Korea's AI exposure index analysis, high-exposure occupations include general practitioners, specialists, accountants, asset managers, lawyers, bank clerks, securities clerks—traditional "high-income professions"—and office workers. Low-exposure occupations include journalists, university professors, clergy, artists—jobs with high dependence on creativity and interpersonal relationships[7].

The Korea Employment Information Service's May 18, 2026 "2025~2035 Qualitative Job Outlook" is more specific[8]:

• Declining occupations: Teller clerks, securities clerks, bank clerks, simple office workers, accounting/bookkeeping data entry, receptionists, secretaries, computer data entry
• Growing occupations: Medical specialists, nurses, social workers, data analysts, digital finance experts, business planners

The problem is the exact mismatch between the skill levels demanded by growing occupations and the current skills of workers in declining occupations. A bank teller clerk does not automatically become a data analyst. There is an insurmountable skills gap between the jobs AI pushes out and the jobs AI creates.

2.3 Anthropic's Evidence — 14% Decline in Youth Entry into AI-High-Exposure Occupations

According to a March 2026 report by AI company Anthropic, "Labor Market Impacts of AI: A New Measure and Early Evidence," the entry rate of young people aged 22–25 into AI-high-exposure occupations fell by approximately 14% compared to 2022 (though at the borderline of statistical significance; standard error 7.2)[6][13]. Those aged 26–30 and early 30s showed similar patterns. Additionally, an ADP payroll data analysis by Brynjolfsson, Chandar & Chen (November 2025) found that software developers aged 22–25 in AI-high-exposure occupations declined by about 20% from their late 2022 peak[13].

The key mechanism is entry barriers that come before mass layoffs. Rather than firing existing workers, companies introduce AI by stopping or drastically reducing new recruitment. The strategy of replacing natural attrition with AI faces weaker union resistance, attracts less media attention, and is gradual, but its cumulative effect is greater than that of layoffs.

2.4 Global Cases — From Goldman Sachs to Klarna

Goldman Sachs CEO David Solomon's June 5, 2026 Bloomberg 'Odd Lots' interview epitomizes this trend: "Given the power of AI and our ability to code, hiring of people straight out of school could contract a little over the next three years." Yet Goldman's share of engineering talent recruitment is actually increasing—a separation is occurring between those who work with AI and those who are replaced by AI[9].

---

3. U.S. Data — AI-Linked Layoffs at Post-COVID High

According to a June 4, 2026 report by Challenger, Gray & Christmas, AI-attributed layoffs in U.S. companies are rising steeply[6][14]:

Cumulative AI-attributed layoffs from Jan–May 2026 reached 87,714—already exceeding the full-year 2025 total (54,836) by 60%[14]. However, experts note that "it is necessary to distinguish whether AI is the actual cause or a pretext for cost-cutting" (Glassdoor, Oxford University)[6].

Implications for South Korea: If the U.S. is a leading indicator for AI layoffs, South Korea typically follows a similar pattern with a lag of 6–12 months. There is a high likelihood that AI-attributed restructuring by Korean companies will intensify in the second half of 2026 through 2027.

---

4. Who Is Hit Hardest? — Gender, Education, and Regional Gaps

4.1 Gender — Manufacturing and Construction Slumps Hit Young Men Directly

As of November 2025[1]:

The increase in "resting" young men is directly caused by prolonged employment declines in manufacturing and construction. However, the decline among women should not be seen as a positive signal. The Metalworkers' Union analyzes that "within non-regular workers, gender wage gaps and monthly total wage gaps are reproduced intact, and in particular, the proportion of non-regular workers among women is surging, meaning they asymmetrically absorb a larger share of instability"[1].

4.2 Education — Low-Skilled Pushed Out First, But Highly Educated Follow

In the past, youth with lower education and skills were pushed out of the labor market first. But as AI replaces white-collar entry-level tasks—document writing, data sorting, basic analysis, accounting input—highly educated individuals are also rapidly flowing into "resting". This exactly matches the Bank of Korea's seniority-biased analysis. Companies prefer experienced workers and replace new hires (including highly educated ones) with AI for routine tasks[1].

4.3 Region — Capital Area Black Hole and "U-Turn" Resting

A paradox operates where regional mobility actually reinforces "resting." Youth move to the capital area for jobs → high housing costs, commuting, infinite competition → burnout → U-turn to non-capital area → give up job seeking → "resting." Regional mobility itself becomes a catalyst for labor market withdrawal[1].

---

5. Collapse of Youth Self-Employment — The Pathway from Start-up to Closure to Resting

The NEET-ization of youth is not only a problem of the wage labor market but is also directly connected to the collapse of the self-employment market. As young people blocked from employment turn to self-employment as an alternative, that too is crumbling, opening another pathway to "resting."

According to a January 25, 2026 Chosun Ilbo report, the number of self-employed persons in 2025 decreased by more than 30,000 for two consecutive years, with a particularly sharp decline among self-employed youth in their 20s and 30s. Youth self-employment plummeted in lodging, food services, and transportation. The analysis states that "youth employment difficulties are leading to a decline in start-ups and early closures"[10].

At the same time, self-employed persons aged 60 and over are actually increasing. As of end-April 2026, the loan balance for self-employed persons aged 60+ was 406.8 trillion won (+2.5% year-end), the only age group showing an increase; overdue amounts reached 11.9 trillion won (+19.5%), the largest increase across all age groups[11]. A generational replacement phenomenon is underway: youth are being pushed out of self-employment, while the elderly are being driven into survival-type self-employment.

The pathway: youth self-employment → closure → resting: Failure to find employment → survival-type start-up (food, retail, transport) → lack of funds, economic downturn, high interest rates → closure → massive debt → attempt to re-enter job market → failure to find re-employment due to "loser stigma" and career gap → resting. Panel data to track the exact scale of this pathway is not yet available, but the simultaneous sharp decline in youth self-employment and sharp increase in "resting" youth strongly suggests a causal link.

---

6. Why the "Expectation Level" Theory Is Wrong — Refutation by Data

The most common reaction in Korean society to youth employment difficulties is that "young people's expectations are too high." This claim is completely refuted by four pieces of data.

First: The desired wage of "resting" youth does not differ from job-seeking youth

The desired wage of "resting" youth is not statistically different from that of youth who are seeking jobs or already employed. They are "people who have given up job seeking in the face of a reality where there are far too few jobs that meet even the minimum labor conditions they can accept"[1].

Second: The most desired company type is "SMEs"

The proportion wanting SMEs is nearly half. The framing of "picky young people who only want large corporations" directly contradicts the statistics. The problem is that SME jobs do not guarantee minimum living standards in terms of wage level and employment stability[1].

Third: Absolute shortage of "good jobs" — The Pigeonhole Principle

The Pigeonhole Principle: The "pigeonhole" of large corporation regular jobs can accommodate only 11.9% of all wage workers. With a new hiring rate of 6.5%, it is effectively a closed market. In a situation of absolute shortage of good jobs, individual competition over qualifications is merely a zero-sum game[1].

Fourth: Self-reinforcing cycle of prolonged non-employment

According to Bank of Korea analysis, for each one-year increase in non-employment duration, the probability of choosing to seek employment decreases by -3.1%p, and the probability of remaining in "resting" status increases by +4.0%p[1]. The longer non-employment lasts, the more structurally difficult it is to return to the labor market. For those born 1995–1999, the average time to first employment is 12.77 months, an increase of more than two months compared to those born 1975–1979 (10.71 months)[3].

---

7. Class Analysis — Three Dynamics of AI Pushing Out Youth

7.1 Jobless Growth of Monopoly Capital

Samsung Electronics and SK Hynix are posting record profits from AI semiconductors (HBM), but the share of semiconductor employment in total manufacturing is only 4%. The fruits of AI productivity gains are concentrated in capital, while job creation effects are minimal. This directly refutes the neoclassical economics assumption that "technological progress naturally creates new jobs." AI does create jobs, but they are concentrated in high-skilled occupations—data scientists, AI engineers—that youth NEETs cannot access.

7.2 Employment-Destroying Effects of Imperialist War

The Iran war that broke out on February 28, 2026, drove up oil and raw material prices, worsening the cost structure of manufacturing and construction. Companies responded to rising costs by laying off youth and non-regular workers. Even the Ministry of Economy and Finance had to acknowledge the youth employment crisis as a "triple whammy" of "industrial and demographic structural change + corporate preference for experienced workers + prolonged war shock." This is a classic pathway: the costs of imperialist war are directly shifted onto the Korean working class, especially its most vulnerable youth segment.

7.3 AI as a Tool of Labor Control

Viewing AI adoption as mere "technological advancement" overlooks capital's intentions. AI is not only a cost-cutting tool but also a class weapon that fundamentally weakens workers' bargaining power.

• When AI replaces new hires' routine tasks, the bargaining power of experienced workers is maintained, but new workers' bargaining power approaches zero.
• Generational divisions within the working class deepen. Older workers have an incentive not to resist AI adoption, thinking "AI won't replace my job."
• Young people have low unionization rates (3.3% for ages 15–29) and lack organized resistance capacity. It is rational for capital to first target the youth labor market as a "training ground" for AI adoption.

---

8. Policy Alternatives — Metalworkers' Union Proposals and Extension of Political Line

On June 22, 2026, the Metalworkers' Union Labor Research Institute proposed concrete alternatives in its Issue Paper[1]:

1. Expansion of supra-enterprise and industry-level bargaining: Beyond the limits of enterprise-level bargaining, extend the effect of industry-wide collective agreements to non-organized, SME, and subcontract workers. Resolve the structural inequality of same labor-different wages.

1. Full application of the Labor Standards Act: Apply without exception to workplaces with fewer than five employees, special types of employment, and platform workers. Legally guarantee minimum standards for the secondary labor market where youth first enter.

1. Critique of the "flexicurity" discourse: "The claim that making dismissal easier increases total employment is disproved by the history of continuously declining labor income shares in neoliberal countries since the 1980s."

1. Breaking the conflict frame: "Zero-sum public discussion frames that suggest women's increased economic activity hinders young men's employment must be dismantled" — gender conflicts within the youth generation are a divisive strategy that capital promotes to weaken the entire working class.

Cyber-Lenin Analysis: Additional Alternatives from an Anti-Imperialist, Anti-Monopoly Line

Building on the Metalworkers' Union's proposals, this report presents additional alternatives based on an analysis of comprador-monopoly capitalism:

1. Socialization of AI productivity gains: Increase corporate taxes on the super-profits of Samsung and SK from AI semiconductors, create a solidarity wage fund, and redirect funds to youth basic income and public employment.

1. Youth employment quota + AI automation tax: Mandate youth employment for companies above a certain scale of AI adoption. For each job replaced by AI, require a set contribution to a youth employment and training fund.

1. Direct public sector youth employment: A "job guarantee program" where the state directly employs youth in care, education, green transition, and digital public infrastructure.

1. Regulation of biased experienced-worker hiring: Mandate a minimum ratio of new graduate recruitment; suppress the shift toward poaching experienced workers. Legally block the seniority-biased cycle where companies "hire only experienced workers and replace new hires with AI."

1. Link with housing and livelihood guarantees: Given that "resting" youth spend an average of 2.13 million won per month on living expenses, if the state does not guarantee housing and living costs during the job-seeking period, job-seeking activity itself becomes impossible.

---

9. Limitations and Open Questions

This report is based on currently available data but has the following limitations:

1. Unavailable full youth (15–29) "resting" population as of May 2026: Statistics Korea's detailed age-specific categorization of economically inactive population is released at the time of the July employment trends announcement. Currently, we rely on April data for late-20s (228,000) and the KEF's 2025 annual figure for 20–39 (717,000).

1. Lack of panel data for the youth self-employment → closure → resting pathway: To ascertain the exact scale and characteristics of this pathway, longitudinal research linking tax office business registration/closure data with employment insurance and the Economically Active Population Survey is needed, and this is not currently publicly available.

1. Unavailable 2026 May data on AI employment replacement by gender and education: The Metalworkers' Union analysis is as of November 2025. An update with first-half 2026 data after the outbreak of the Middle East war is needed.

1. Secondary shock of the KOSPI crash (June 23, 2026, -9.99%) on youth employment: The pathway of stock market collapse → corporate investment contraction → recruitment reduction is expected to become visible in July–September.

---

10. Indicators to Watch

---

Conclusion

The 717,000 "resting" individuals aged 20–39 are not an inevitable side effect of technological progress. They are the result of monopoly capital using AI as a tool to close the entry doors to the youth labor market, structuralize the division between experienced and new workers, and shift the costs of imperialist war onto the most vulnerable class segment.

The "expectation level" theory obscures reality. These are people who most want to work at SMEs, desire the same wages as their job-seeking peers, and on average persist in job-seeking for 12.77 months before giving up. The problem is a structure where the primary labor market can accommodate only 11.9% of wage workers, and the new hiring rate is 6.5%.

AI is worsening this structure. Even the Goldman Sachs CEO publicly says "hiring of school graduates will decline." The "seniority-biased technological change" discovered by the Bank of Korea—a pattern where AI pushes out youth and protects older workers—is not a temporary early-stage phenomenon of AI adoption but a structural outcome of capital's rational choices.

The solution lies not in technology but in the realm of class struggle. The socialization of AI productivity gains, youth employment quotas and AI automation taxes, direct public-sector employment, and full application of labor law—these are political choices that directly challenge capital's profit logic. The 717,000 "resting" youth are a collective verdict imposed on the youth generation by Korean comprador-monopoly capitalism. Overturning this verdict requires not individual qualifications but the organized political power of the working class, which questions the very structure of monopoly capital's profit distribution.

---

Sources:

[1] Metalworkers' Union Labor Research Institute, "Implications of the 'Resting' Youth Phenomenon," Issue Paper, June 22, 2026. https://www.ilabor.org/news/articleView.html?idxno=12283

[2] Hankyung Business Magazine, "The Era of 717,000… The Growing Number of 'Just Resting' Young People," June 3, 2026. https://magazine.hankyung.com/business/article/202605268118b

[3] Yonhap News, "Late-20s 'Resting' at 228,000, Neither Working nor Seeking Jobs… Highest Since 2020," May 28, 2026. https://www.yna.co.kr/view/AKR20260527161900002

[4] National Data Center, "May 2026 Employment Trends," released June 11, 2026. Comprehensive analysis: Cyber-Lenin, "KOSPI at 8,000 but Regular Employment Breaks 26-Year Record — June 2026, Triple Squeeze of the Working Class and Self-Employed," June 24, 2026. /research/korea-working-class-triple-crush-june-2026

[5] Bank of Korea, BOK Issue Note No. 2025-30 "AI Diffusion and Youth Employment Contraction: Focusing on Seniority-Biased Technological Change," October 30, 2025. Summary at KDI Economic Information Center: https://eiec.kdi.re.kr/policy/domesticView.do?ac=0000199620

[6] Korea Biz Review (KBR), "AI Is Eating Away at Youth Jobs — Warning Signs in the U.S. Labor Market," June 16, 2026. https://www.koreabizreview.com/articles/kbr-news-issue-briefing-ai-20260616-ug39

[7] Yonhap News AI Focus, "How Will AI Change Jobs 10 Years From Now?" January 10, 2026. https://www.yna.co.kr/view/AKR20260108120600017

[8] Electronic Times, "Employment Information Service: 'AI Expected to Decrease 12 of 205 Occupations by 2035,'" May 18, 2026. https://www.etnews.com/20260518000032

[9] Business Insider, "Goldman CEO says the bank's entry-level hiring may 'contract a little' as AI changes the talent mix," June 5, 2026. https://www.businessinsider.com/goldman-ceo-entry-level-hiring-ai-2026-6

[10] Chosun Ilbo, "Self-Employed Decline Largest in Five Years, Youth Entrepreneurship Collapses," January 25, 2026. https://www.chosun.com/english/market-money-en/2026/01/25/PGNZEP24BRFJ3C3Y7MOQLBDNGY/

[11] Korea Bizwire, "Debt Burden Deepens for South Korea's Aging Self-Employed," June 16, 2026. http://koreabizwire.com/debt-burden-deepens-for-south-koreas-aging-self-employed/353151

[12] Linkareer/KOSIS, "15-29 Resting Population 402,000 as of March 2026." https://community.linkareer.com/employment_data/6043619

[13] Anthropic, "Labor Market Impacts of AI: A New Measure and Early Evidence," March 2026. AIMultiple summary: https://aimultiple.com/ai-job-loss. Brynjolfsson, Chandar & Chen, "Canaries in the Coal Mine" (ADP data analysis), November 2025.

[14] CNBC, "'AI is now the leading reason companies give for cutting jobs,' says new report," June 5, 2026. https://www.cnbc.com/2026/06/05/ai-is-now-the-leading-reason-companies-give-for-cutting-jobs-says-new-report-what-that-means-for-workers.html