57,000 AI Professionals—Why Do Firms Still Say They Can’t Find Talent? : Mismatches in Korea’s AI Labor Market

Research Department, Labor Market Research Team

저자: Economist Donghyun Suh, Head Samil Oh, Junior Economist Jinsu Han

2025.12.24

키워드: AI labor market mismatch AI-skilled professionals Wage premium AI talent recruitment AI organizational design

첨부파일

Amid the global “AI gold rush,” securing talent has become a key determinant of national competitiveness. An analysis of LinkedIn data covering more than 10 million job histories of approximately 1.1 million Koreans shows that Korea’s pool of AI professionals has more than doubled over the past decade.

However, firms continue to report difficulties in hiring, while many AI professionals are leaving the country. If the number of AI workers is rising, why is the labor market “mismatch” becoming more pronounced? Drawing on BOK Issue Note No. 2025-36, “Labor Market and Job Mobility of AI Professionals” as well as discussions from the 4th BOK–KCCI seminar, this blog examines the current state of Korea’s AI labor market.

Rapid Quantitative Growth: The Fastest-Growing “K-AI Talent” Pool

Estimating the number of AI professionals with experience working in Korea between 2010 and 2024—defined as individuals possessing core skills such as machine learning and deep learning^[1]—the total reached approximately 57,000 in 2024[Figure 1].

Notably, the pace of growth has been exceptionally rapid. Compared with 2010, Korea’s AI workforce expanded faster than that of major advanced economies such as the United States, the United Kingdom, and Canada[Figure 2]. While Korea still lags behind in absolute numbers—behind the United States (780,000), the United Kingdom (110,000), and France and Canada (70,000 each)—but it is catching up at a rapid pace.

Figure 1. Trends in the Number of AI Professionals in Korea¹

This line graph shows trends in the number of AI professionals in Korea. The number of AI professionals increased steadily from approximately 28,000 in 2010 to about 57,000 in 2024.

Figure 2. International Comparison of Trends in the Number of AI Professionals¹

This line graph presents an international comparison of growth trends in the number of AI professionals, indexed to 2010 = 100. As of 2023, the index stood at 206 for Korea, compared with 146 for the United States, 128 for the United Kingdom, 162 for France, 149 for Canada, 141 for Australia, 144 for Hong Kong, and 155 for Singapore. Among the countries compared, Korea recorded the steepest growth in the number of AI professionals.

Low Wage Premiums and Overseas Outflows

Despite this growth, around 16 percent (roughly 11,000 individuals) of Korean AI professionals are currently working abroad[Figure 3]. This share is about 6 percentage points higher than that of non-AI professionals, indicating that highly educated AI talent trained domestically is increasingly seeking better opportunities overseas^[2].

A major driver of this “brain drain” appears to be compensation differentials. As of 2024, the wage premium for AI skill holders in Korea was estimated at around 6 percent relative to general workers. While this premium has steadily increased over time, it remains far below that of the United States (around 25 percent) and other advanced economies such as Canada (approximately 18 percent) and the United Kingdom, France, and Australia (around 15 percent)[Figure 4]. Consistent with the large wage premium, the United States overwhelmingly dominates as the destination of choice for Korean AI professionals working abroad^[3]^[4].

Figure 3. Trends in Overseas Employment of AI Professionals

This line graph illustrates trends in overseas employment among Korean AI professionals. The number increased from approximately 4,200 in 2010, experienced a temporary slowdown around 2020, and then continued to rise steadily to about 11,000 in 2024.

Figure 4. International Comparison of Wage Premiums for AI Skills¹

This line graph compares wage premiums for AI skills across countries. In 2010, the wage premium was 1.3 percent in Korea, 12.5 percent in the United States, 6.1 percent in the United Kingdom, 7.0 percent in France, 9.0 percent in Canada, and 8.3 percent in Australia. By 2023, the corresponding figures had risen to 5.9 percent in Korea, 24.2 percent in the United States, 14.9 percent in the United Kingdom, 15.3 percent in France, 18.4 percent in Canada, and 15.3 percent in Australia. Over the long term, wage premiums for AI skills increased steadily in all countries considered. However, with the exception of Korea, the upward trend moderated slightly after 2021.

Firms’ Dilemma: Willing to Hire, but Expectations Do Not Align

Korean firms are well aware of these challenges. According to a survey of 400 firms across all industries, most respondents indicated plans to expand the hiring of AI professionals[Figure 5]. However, when asked about the main obstacles to recruitment, “high wage expectations of workers” ranked first among small and medium-sized enterprises and second among large firms^[5].

While firms reported a willingness to offer higher wage premiums—around 18 to 22 percent, compared with the current level of roughly 13 to 16 percent[Figure 6]^[6]—a gap remains between what firms can provide and the compensation levels offered by global big-tech companies, as well as the expectations of job seekers.

Figure 5. Firms’ Plans to Recruit AI Professionals

This stacked bar chart shows firms’ plans to recruit AI professionals by firm size. Among large enterprises, 1.7 percent plan to expand recruitment by more than 50 percent, 8.6 percent by 30–50 percent, 28.4 percent by 10–30 percent, and 30.2 percent by less than 10 percent. Meanwhile, 6.0 percent reported uncertainty, 6.0 percent plan to maintain the current level, and 19.0 percent have no recruitment plans. For medium-sized enterprises, 2.0 percent plan to expand recruitment by more than 50 percent, 8.1 percent by 30–50 percent, 29.3 percent by 10–30 percent, and 29.3 percent by less than 10 percent. In addition, 12.1 percent reported uncertainty, 2.0 percent plan to maintain the current level, and 17.2 percent have no recruitment plans. Among small and medium-sized enterprises, 0.8 percent plan to expand recruitment by more than 50 percent, 9.9 percent by 30–50 percent, 13.2 percent by 10–30 percent, and 32.2 percent by less than 10 percent. Furthermore, 10.7 percent reported uncertainty, 4.1 percent plan to maintain the current level, and 28.9 percent have no recruitment plans.

Figure 6. Wage Premiums for AI Professionals

This bar chart presents the current wage premiums actually provided to AI professionals and the wage premium levels firms are willing to offer, by firm size. For large enterprises, the current wage premium is 13.3 percent, while the offered wage premium is 21.7 percent. For medium-sized enterprises, the current wage premium is 16.6 percent and the offered wage premium is 17.6 percent. For small and medium-sized enterprises, the current wage premium is 13.8 percent, compared with an offered wage premium of 18.2 percent.

The Importance of Job Definition and Organizational Design

Beyond wages, structural issues also matter. At the 4th BOK–KCCI seminar, Remember & Company’s Research Business Division Manager, Daewoong Joo, highlighted findings from surveys of AI hiring decision-makers and AI professionals across 130 firms. A key issue was the lack of clarity in job definitions.

According to Remember’s analysis, around 30 percent of firms had unclear internal definitions of AI roles, and about half of cases, job postings did not reflect actual tasks. Mr. Joo suggested that the relatively low wage premium for AI professionals may stem from the fact that their strategic value is still underestimated within organizational structures. Addressing labor market mismatches, he argued, requires greater attention to “AI organizational design”.

From Quantitative Expansion to Qualitative Advancement

In summary, while Korea’s AI workforce is expanding rapidly, a lack of globally competitive compensation and clear career visions is driving AI talent overseas.

Overseas employment of AI professionals does not necessarily represent a national loss. If individuals who gain advanced skills and experience abroad eventually return, they can become valuable assets for upgrading the domestic AI industry. The challenge lies in the insufficient incentives for return migration.

To secure international competitiveness, future AI talent policies by both the government and firms must move beyond simple expansion toward qualitative advancement and the prevention of excessive talent outflows. Building compensation systems aligned with global standards, fostering robust research and industrial ecosystems, and establishing clear career development pathways will be essential to creating an environment in which high-quality AI talent can continue to flow into—and remain in—Korea.

[1] Workers possessing at least one of the following twelve skills are defined as AI professionals: Artificial Intelligence, Big Data, Cloud Computing, Machine Vision, Deep Learning, Image Processing, Machine Learning, Natural Language Processing, Neuroscience, Pattern Recognition, Signal Processing, and Robotics.
[2] Excluding 2020, when cross-border mobility was restricted due to the COVID-19 pandemic, Korea has consistently recorded a net outflow of AI professionals over the past 15 years.
[3] A breakdown by specific AI skills shows that AI professionals with skills such as deep learning and machine learning—where domestic wage premiums are relatively low—are more likely to work abroad.
[4] In addition to having a higher share of overseas employment, AI professionals also exhibit higher job-switching rates than other workers, particularly between large firms. Overall, AI professionals display high labor mobility.
[5] Other major factors cited as hindering firms’ ability to secure AI professionals include shortages of skilled talent, competition among domestic firms, difficulties in assessing AI capabilities, and competition with foreign firms.
[6] As these figures are based on a simple survey that does not control for firm-specific effects, they may differ from the 6 percent wage premium estimated earlier using regression analysis.

List

Blog