1. 285,000 Youth Jobs Vanished, 94% Concentrated in One Area

Recent youth employment slumps display characteristics distinct from the past. A Bank of Korea (BOK) Issue Note released on August 18, 2026, titled "Youth Employment Contraction: Is AI to Blame? Changing Career Ladders and Policy Tasks," analyzed shifts in youth employment from June 2022 to June 2026.

During this period, jobs for young people aged 15 to 29 dropped by 285,000. Notably, 268,000—or 94%—of the lost jobs were concentrated in sectors heavily exposed to AI. Youth employment fell by 31.4% in the information services sector, 27.4% in publishing, and 16.6% in computer programming, system integration, and management. These industries are relatively more susceptible to the impact of AI and digital technologies.

Even more striking is the disparity across age groups. Over the same timeframe, jobs for individuals in their 50s rose by 230,000. Among these, 173,000 jobs (75.2%) were added within industries highly exposed to AI. Within the exact same AI-exposed sectors, youth employment shrank while middle-aged employment expanded.

This alone does not justify the conclusion that "AI eliminated 285,000 youth jobs." Other factors, such as the normalization of post-pandemic over-hiring in the tech sector, corporate preference for experienced hires, and the spread of remote work, may have also played a role. Nevertheless, the heavy concentration of youth job losses specifically in AI-exposed industries is a new phenomenon that demands explanation.

2. Why AI Is Harsher on the Youth Than Middle-Aged Workers

The clue lies in the distinct roles junior and experienced workers play within an organization.

Junior staff typically begin their careers by organizing data, summarizing documents, creating report drafts, and writing basic code. These tasks are less about demanding advanced expertise and more about learning the organization and workflow through repetitive execution. These are also the exact domains generative AI is penetrating most rapidly. Experienced workers are different. They grasp organizational context, coordinate relationships with clients, handle exceptional scenarios, and make decisions among multiple stakeholders. These responsibilities rely heavily on "tacit knowledge" and experience that cannot be easily codified into manuals.

Therefore, when AI is introduced, rather than replacing the entirety of an experienced worker's responsibilities, companies are far more likely to first automate the tasks traditionally handled by entry-level employees.

Interesting changes are also emerging across educational attainment levels. Before the release of ChatGPT, from January 2019 to October 2022, the average unemployment rates for college-educated youth and youth with a junior college degree or less were nearly identical at 8.2% and 8.0%, respectively. However, from the release of ChatGPT through June 2026, the average unemployment rate stood at 7.0% for college graduates and 5.4% for those with a junior college degree or less.

While the youth unemployment rate for college graduates itself fell from 8.2% to 7.0%, their relative standing shifted. The unemployment gap between the two groups widened from just 0.2 percentage points before ChatGPT to 1.6 percentage points afterward. This is a noteworthy development consistent with the characteristic of generative AI impacting knowledge workers and white-collar roles before manual labor.

3. Entry Doors Narrow While Exit Doors Widen

The Bank of Korea also analyzed the inflow of young people into new jobs and their outflow into unemployment within AI-exposed industries. Comparing pre-pandemic figures with recent data, the monthly average number of youth entering high-AI-exposure sectors fell from 32,600 to 29,100, a decrease of approximately 11%. Conversely, the monthly average number of young workers leaving these sectors into unemployment rose from 3,700 to 4,900, an increase of about 32%. The entrance door has narrowed, while the exit door has widened.

The anxiety felt by young jobseekers and employees reflects this trend. In a survey conducted by the Korea Economic Association (KEA) involving 1,000 employed and job-seeking youths aged 19 to 34, 63.8% of jobseekers responded that their desired job functions are highly likely to be replaced or reduced by AI within the next five years. Among employed youth, 49.4% also viewed the reduction or replacement of their current roles as likely. For workers in high-AI-exposure roles, 58.4% foresaw high replacement potential.

A particularly notable response was that 65.6% agreed that "as AI replaces entry-level tasks, opportunities to develop into highly skilled talent will diminish." This shows that young people are not simply afraid that "AI will steal my job."

4. What Happens When the 'Career Ladder' Vanishes Instead of Jobs

In the past, the simple tasks assigned to new employees might have seemed inefficient. A task taking a senior worker one hour might take a junior employee three to four hours, with drafts full of errors requiring revisions. From an organization's perspective, however, this was not merely low-productivity labor. By organizing data, they learned the industry; by drafting reports, they learned logical structuring; and by receiving feedback from seniors, they internalized decision-making criteria. Today's mundane tasks served as the training ground to produce the skilled workers of tomorrow.

Analyzing the potential impact of AI on young, highly educated workers, the International Monetary Fund (IMF) also highlighted the risk that they could struggle to secure "stepping-stone jobs" essential for building a career.

From a company's standpoint, eliminating tasks that are easiest to replace with AI may seem rational. But when all firms act this way, systemic problems emerge. If a company that once hired 10 entry-level workers to train them now hires only 3 with the help of AI, immediate corporate productivity might rise. Yet, if multiple companies adopt this strategy simultaneously, the total number of entry points for learning work across society will shrink.

In 5 to 10 years, another question arises: Where will companies that did not nurture junior talent find future experienced professionals? This is why experts point out that the real risk of the AI era for youth employment may go beyond simple "job losses" to the dismantling of the career ladder.

5. Yet, AI Helps Novices the Most

An interesting paradox emerges: while AI threatens entry-level jobs, junior workers may also be the ones who benefit the most from AI assistance.

Research supports this notion. A study tracking productivity changes before and after the introduction of generative AI tools among 5,179 customer support agents at a large U.S. software company (Erik Brynjolfsson of Stanford University, Danielle Li and Lindsey Raymond of MIT, NBER Working Paper No. 31161) demonstrated striking results.

Employees using AI saw an average productivity increase of 14%. For novice and low-skilled workers, productivity improved by 34%, whereas the effect on experienced workers was relatively small. The researchers interpreted that AI acted as a "digital coach," transferring the know-how of experienced personnel to beginners.

This study does not prove that AI reduces junior hiring. Rather, it demonstrates how rapidly AI can accelerate the learning and productivity of novices.

This raises a new question: If one junior employee using AI can perform the work of two or three people from the past, will companies hire the same number of entry-level workers as before?

For individual young workers, AI is a powerful tool to enhance personal productivity. However, just because every young worker improves productivity with AI does not mean overall employment opportunities for the youth will expand. What benefits the individual does not necessarily benefit the labor market as a whole.

In fact, in the KEA survey, 64.7% of youth responded that AI helps enhance their expertise and career competitiveness, with that figure rising to 74.4% among those in high-AI-exposure roles. At the same time, 67.7% expressed concern that disparities in AI proficiency would widen income and career divides within the younger generation.

6. What Matters Is Not Whether AI Is Used, but 'How It Is Used'

The Bank of Korea's analysis offers an important clue: youth employment dropped more sharply in industries where AI is utilized primarily for "automation," replacing human tasks entirely. In contrast, sectors where AI was leveraged for "augmentation"—assisting human work and enhancing productivity—did not exhibit the same pattern of steep employment declines.

Ultimately, what matters may not be the existence of AI itself, but what companies use it for. A firm can use AI to eliminate the drafting tasks junior workers used to perform, or it can have junior workers verify, edit, and apply AI-generated drafts to real-world scenarios. While the former focuses on cost-cutting, the latter designs a new training process utilizing AI.

Young people are demanding similar solutions. In the KEA survey, the top policy measures selected were "strengthening support for companies hiring new youth talent" (23.8%) and "expanding opportunities for entry-level work experience" (23.7%).

Youth employment policies in the AI era must move beyond simply measuring "how many people were hired." The focus should shift to how to provide experiences in learning real work, exercising judgment, and taking responsibility while using AI. This approach creates incentives for companies to design new entry-level roles and offers young workers the experience needed to advance to higher-level work using AI.

Whenever new technologies emerged after the Industrial Revolution, fears that "machines will take human jobs" recurred. AI may ultimately follow a similar trajectory, with some occupations disappearing and new ones emerging as the labor market adapts.

However, generative AI possesses a distinct characteristic compared to previous technological shifts: it is rapidly encroaching on the very domain where humans first learn how to work. The process of finding data, organizing it, drafting, making mistakes, and receiving revisions appeared inefficient, making it the easiest target for AI replacement.

Yet, that inefficiency may have been the necessary investment required to transform a person into a skilled professional. This is why companies must distinguish whether what they are eliminating with AI is merely simple work, or the vital training process that produces tomorrow's skilled labor.

The question surrounding youth employment in the AI era must also shift: from "How many jobs will AI destroy?" to "Where will junior employees learn to work in the age of AI?" If AI removes the first step of the career ladder, building new steps is ultimately not the role of technology, but the responsibility of businesses and society.


Deep Dive Q&A
Q1. Does this mean 285,000 youth jobs disappeared because of AI?

A: No. The Bank of Korea's analysis shows that 94% of the 285,000 youth jobs lost between June 2022 and June 2026 were concentrated in industries with high AI exposure. This alone does not conclusively prove AI as the direct cause of the job losses. Other factors, such as the normalization of pandemic-era over-hiring in the tech sector and corporate preference for experienced workers, may have also had an effect. However, the heavy concentration of youth job losses in AI-exposed sectors and the larger declines in automation-heavy industries highlight the need to pay close attention to AI's impact on youth employment.

Q2. If AI boosts productivity, isn't it actually beneficial for young workers?

A: For individuals, that is very likely. In an empirical study of customer support agents, AI usage increased average productivity by 14%, and by 34% for novice and low-skilled workers. The problem is that individual productivity gains do not translate to an increase in overall hiring for youth. If companies determine they can handle the same workload with fewer entry-level hires by utilizing AI, overall new hiring could decrease. Thus, the crucial question is whether AI is used to replace junior workers or to help them become proficient faster.

Q3. What kind of work should entry-level employees do moving forward?

A: Rather than avoiding what AI does well, entry-level roles should be redesigned around evaluating and taking responsibility for AI outputs. For instance, while junior staff previously wrote report drafts from scratch, in the future they could verify facts in AI-generated drafts, identify errors and biases, adapt content to organizational context, and apply it to real-world situations. In this scenario, required skill sets will also shift. Rather than basic document drafting or simple information searches, capabilities such as verification, critical thinking, contextual comprehension, communication, and AI utilization will likely become far more essential.

Q4. In the end, won't young workers who are proficient with AI have an advantage?

A: That is likely. In the KEA survey, 64.7% of youth—and 74.4% in high-AI-exposure roles—responded that AI helps improve their expertise and career competitiveness. However, an issue of disparity remains. In the same survey, 67.7% worried that differences in AI utilization capability would widen income and career divides within the youth demographic. Consequently, the key issue in the future labor market may not be merely "those replaced by AI versus those who are not," but rather the divide between those who can use AI to move to higher-level work and those who cannot.