Hyundai Motor Group and Boston Dynamics demonstrated the capabilities of their new humanoid robot, Atlas, in a factory setting. (courtesy Hyundai Motor)
Min-su, a software developer with three years of experience under his belt, recently started working at a well-known Korean IT conglomerate. “When I first started working in 2023, your aptitude for writing code was what was most important, but recently, companies are favoring those who utilize AI to write code for them,” he said, noting the abrupt change he felt during his interview process.
Around the office, colleagues whisper about needing to somehow show how they’ve boosted productivity with AI, lest they get bad marks on their next performance review. Min-su also uses AI in around 95% of his tasks.
In 2022, the total number of new hires — including both newcomers and experienced professionals — across all of the company’s subsidiaries was around 600, but this figure dropped sharply to 259 in 2024.
Min-su — a pseudonym granted for the sake of this article, as are other given names that appear hereafter — is certain that AI was the biggest factor in this drop. Min-su, who is in his early 30s, is the youngest in his team of 10.
This is not a phenomenon unique to Min-su’s current company. According to figures from the Software Policy & Research Institute, the share of Korean developers with less than three years’ experience dropped from 26.9% in 2022 to 20.7% in 2024. This was the only experience bracket that saw a decrease in proportion.
Since ChatGPT’s launch in 2022, generative AI has gradually crept into our daily lives. During this time, it has also triggered a quiet trend of corporate downsizing.
As employers continue to favor those with experience, AI is now a key factor in many companies’ decision to hire fewer people for entry-level positions. AI-triggered job insecurity is becoming increasingly evident among vulnerable groups in the job market, especially young Koreans.
According to a report on AI and labor published by the Korean Labor Institute, 58% of the 500 office and administrative workers surveyed expressed concerns about losing their jobs to AI within the next 10 years.
Anxieties were highest among those with the lowest seniority (ranked “daeri” or lower), at 36.4% — 10 percentage points higher than that of managers at the department head level and above. In terms of age, those in their 20s (35.2%) were more concerned than those in their 50s (27.3%).
To young people who have yet to enter the labor market or who have minimal experience, the difference is palpable. Ji-eun, a 23-year-old majoring in economics, is approaching graduation while seeking a job. Last year, she worked as an intern at a startup that manufactured fashion products and cosmetics. Even before setting foot in the labor market, she sensed the threat posed by AI.
“The CEO assigned a market research task to me. I spent two days making an Excel sheet to arrange all the data. My boss subscribes to a generative AI service, and within five minutes, it spat out data that was better than what my excel sheet offered,” Ji-eun said.
“I worked hard and researched as much as I could, but the AI’s results were so much better, they were incomparable,” she added.
The CEO appeared to trust the AI service more than Ji-eun’s abilities. She started second-guessing things.
“I think I need to do something to reduce the risk of being replaced by AI, but I’m not sure what that is,” she lamented.
Entry-level jobs gobbled up by AI: Paperwork, summaries, research, etc.
AI is already doing a substantial portion of the tasks normally assigned to new employees.
The Korea Labor Institute’s survey found that 30.2% of respondents said that key tasks that constituted a new employee’s standard training — such as writing summaries and first drafts of reports, research, etc. — are now being delegated to AI.
Some observers say that AI and robots are taking away opportunities for beginners to observe, aid and learn from those with more experience and experts, kicking away the ladder of skill-building.
When asked about the kinds of tasks that AI is frequently used for, KLI survey respondents replied “drafting documents and reports” (68.8%), “internet searches for information” (61.8%), “data analysis” (46%), “reviewing reports” (37%), “translation” (36.6%), and “writing meeting minutes” (32.8%).
In the same survey, 23.4% of respondents said that the departments they worked for had reduced recruitment or were planning to cut back on hiring.
“It’s not only us. Other conglomerates are also delegating research and other tasks that had been assigned to inexperienced employees to AI,” said Jeong-hee, the head of HR and recruitment at a major corporation’s holding firm. “Even if we hire people for entry-level positions, there’s often nothing for them to do, so people in the field ask us to hire people for senior positions.”
“Going forward, companies will hire just enough new employees to keep the company going, no more,” she predicted.
This appears to be borne out in figures. The Bank of Korea’s analysis of employment fluctuations for different age cohorts over three years (July 2022 to July 2025) shows that 211,000 fewer jobs were offered to young people, with 208,000 of them were in fields above the 50th percentile of exposure to AI.
Since the launch of ChatGPT, youth employment in fields like computer programming, system integration and management, publishing, professional services, and information services has decreased.
Conversely, workers in their 50s were offered 209,000 more jobs; 146,000 of them were in fields with high exposure to AI.
“Young employees are assigned to perform routine, standardized tasks, which can usually be performed by AI. Senior positions require ample experience related to understanding professional context or managing an organization — skills and knowledge that are currently difficult for AI to replace,” the central bank noted in its analysis.
However, even if a company hires new employees, hiring standards have gone up, adding to young jobseekers’ troubles.
Su-ji, a 24-year-old Chinese language and literature major, was shocked when she scrolled through the requirements for a corporate strategy internship at a major company, only to find that it required a Level 6 in the HSK examination, a test of proficiency in Chinese.
“That’s higher than the requirement for a linguistic specialist position in the National Intelligence Service,” she said, adding that companies are demanding more of applicants.
After working as a developer for a year, Min-hyeok, 28, is currently on the hunt for a new job.
“Even in 2024, around 20% of my applications would pass the initial screening phase. I’ve applied to around 20 places in the last five months; not a single one replied,” he said.
“To boost my resume, I’ve registered for an AI course,” he added.
Yun Seung-gyu, an employment consultant, has also seen the shift in hiring practices.
“Employers are asking questions they used to ask experienced professionals during interviews for new, inexperienced employees,” he said. “Going beyond simple tasks, companies are demanding experience in solving problems using AI, or the requisite ability.”
AI in the assembly line chases out over 200 factory workers at a subcontractor
But young people aren’t the only ones affected. For the past two to three years, firms have been finding tasks they can delegate to generative AI, and have even created departments devoted to automation.
A firm that specializes in producing multi-layer ceramic capacitors (MLCCs; used to stabilize the flow of electricity in electronics) for semiconductors started applying AI to existing equipment in 2019, and has steadily pursued the automation of various parts of the production phase.
AI is now being used in a process that once required hundreds of workers at a subcontractor to inspect each product with microscopes. The AI is fed photographic training data about acceptable product quality and defective products; it is also given inputs of supplementary data to enhance its ability to detect defects.
The introduction of AI to the assembly line has resulted in over 200 factory workers leaving the subcontractor over the past few years.
“By applying AI, we have reduced the number of employees required for managing machine sorters from one person per five sorters to one person per 15 sorters,” explained Yeong-ho, who belongs to the firm’s automation division.
“We’ve even applied AI to the last production phase for circuit boards, which requires a person to directly confirm the results, thereby not only increasing detection of product defects but reducing the frequency of personal observation,” he added.
Even at a managerial position, Yeong-ho admitted that the speed of technological development in AI scares him.
“When agentic AI [which sets objectives and executes actions under minimal supervision] meets physical AI, the current manufacturing workforce may be cut in half,” he predicted.
Freelancers, who often fall outside the protections of labor laws, are also being hit by AI.
Yoon-jin, a 34-year-old freelance video editor with nearly eight years of experience, said, “Before AI arrived on the scene, I got paid two to three million won for a one-minute video that would take two to three weeks to make.”
“Editors with just one to two years of experience can use AI to make videos that are similar to those made by veterans with 20 years of experience,” they added, “which has driven down editors’ rates.”
Concerns about increasing inequality in the labor market
Experts warn about the potential for AI to exacerbate inequality in the labor market. Even the Korea Labor Institute survey indicated greater concerns about people in lower income brackets and in smaller companies losing their jobs.
“Compared to their peers at larger corporations, employees at SMEs are more acutely feeling the double whammy of insecurity from not only the short-term risk of job loss due to business closures stemming from declining corporate competitiveness, but also the fact that they may never even get the chance to build up AI skills, leaving them at risk of being pushed out of the labor market in the long term,” the institute analyzed.
Kim Jong-jin, the director of the Korea Worker Institute, warned that already vulnerable workers may be the first to be replaced.
“Because AI is automating tasks rather than whole jobs, freelancers and subcontracted workers who lack legal and union protections are at risk of becoming the first to be replaced by AI,” Kim said.
There are also rising concerns that as companies ramp up AI use and stop hiring as many entry-level workers, it could lead to a dearth of human capital down the line.
“While at the individual company level it may be the rational choice to not hire new workers and instead boost productivity by having existing employees use AI, at the national level, this means a massive loss of human capital,” assessed Cho Seong-jae, a senior fellow at the Korea Labor Institute.
In short, the disappearance of opportunities for young workers to develop their skills could mean that not enough people are being trained to take on mainstay positions at companies down the line.
By Nam Ji-hyeon, staff reporter
Please direct questions or comments to [english@hani.co.kr]