However, according to a Stanford University study, South Korea's net inflow of AI talent stands at -0.36 per 10,000 residents, ranking among the lowest globally. The reality is that the rate of departure outpaced the rate of cultivation.
The Uncomfortable Reality of a Nation Declaring Itself an 'AI Big 3' Power
In January 2026, the "Framework Act on Artificial Intelligence" (AI Framework Act) went into full effect in South Korea.
The year 2026 is projected to mark the inaugural year of South Korea's AI industry institutional transition, with the AI Framework Act fully implemented starting January 22. The government has set a goal of leaping into the world's top 3 AI powerhouses and is consecutively rolling out cross-ministerial talent cultivation measures. The institutional framework is taking shape, but the numbers remain cold.
According to a Stanford University study, South Korea's net inflow of AI talent is -0.36 per 10,000 residents (representing a net outflow), placing it at the very bottom globally. Even Seoul National University failed to fill 75% of its admissions quota for science and engineering graduate programs in the spring semester of 2025. The gap between declarations and reality is wide. This report closely examines that gap.
1. Where Does South Korea's AI Talent Cultivation Infrastructure Stand Today?
Establishment of the Institutional Framework
Since 2022, the South Korean government has pursued cross-ministerial policies with the goal of cultivating "1 million digital talents." To support the expansion of digital education opportunities and competency enhancement for all citizens, the government formulated the "Comprehensive Plan for Digital Talent Cultivation," aiming to train a total of 1 million digital professionals from 2022 through 2026.
In November 2025, the Ministry of Education unveiled the "Artificial Intelligence (AI) Talent Cultivation Plan for All." This plan encompasses an AI education system spanning the entire life cycle, from elementary school to lifelong education. Specifically, it includes expanding AI and SW-specialized programs for science high schools and gifted schools from 14 institutions in 2025 to 27 in 2026, and introducing a bachelor's-master's-doctoral fast track to shorten the traditional 8-year doctoral acquisition process to 5.5 years. At the graduate level, the expansion of quotas for AI and AX (AI Transformation) graduate schools and the introduction of Brain Korea 21 (BK21) AX research groups (3 groups with 4.2 billion KRW in 2026) are also being pursued.
In 2026, the Seoul Metropolitan Government also stepped in. It announced the full-scale promotion of AI and science-engineering talent cultivation, investing 15.1 billion KRW annually and approximately 60 billion KRW over four years to foster 2,000 professionals each year. It will select 11 universities, providing 600 million KRW to the top-ranked university and 500 million KRW to each of the others.
Meaningful plans were also announced regarding the R&D budget. The government proposed a policy to maintain the total R&D budget at around 5% of total government expenditures over a five-year period. The government's 2026 budget proposal reflects a level of approximately 4.8%.
The national R&D project budget allocation plan includes support for top-tier emerging AI researchers (34.3 billion KRW), generative AI leading talent cultivation (27 billion KRW), artificial intelligence convergence innovation talent cultivation (21 billion KRW), attracting top-tier overseas AI talent (10 billion KRW), and artificial intelligence research hub projects (10 billion KRW).
Efforts are also underway to strengthen linkages between universities and corporations. The foundation for in-house corporate graduate schools where companies directly cultivate AI talent is being strengthened, and the introduction of an (tentatively named) industrial degree system—where degrees are earned based on R&D achievements—is being pursued. In 2026, 10 universities will be newly selected for the cutting-edge industry talent cultivation bootcamp in the AI sector.
On the surface, a substantial policy system is being built. However, institutional completeness does not automatically equate to field performance.
2. South Korea's Position Through Quantitative Indicators: The Discrepancy Between 6th and 11th Place
South Korea's standing in global AI competitiveness indicators varies significantly depending on the metric used.
South Korea possesses the world's 6th highest level of AI competitiveness (Tortoise Intelligence, The Global AI Index, 2023), receiving favorable evaluations in infrastructure, development, government policy, and operating scale indices. In terms of technological level, compared to the United States (100%), South Korea's technical proficiency rose from 81.6% in 2018 to 88.9% in 2022, and the development speed of application-stage AI technologies was found to be more than twice as fast as major countries.
However, when applying the practical yardstick of private investment, the story changes. As of 2024, private-sector AI investments in the U.S. reached 109.08 billion USD (approx. 161 trillion KRW), which is 11.7 times larger than second-place China (9.29 billion USD) and 24.1 times larger than third-place the UK (4.52 billion USD). Conversely, South Korea's investment scale ranking dropped from 9th to 11th year-over-year. Israel and Singapore were cited as the countries with the highest AI talent concentration, while South Korea lingered around the 10th tier.
The gap is also prominent in AI model development performance. According to the Stanford Institute for Human-Centered Artificial Intelligence (HAI) "AI Index Report 2025," the number of notable AI models released by U.S. companies in 2024 was 40, compared to 15 for China. South Korea, along with Canada and Israel, recorded only 1.
A substantial gap also exists in the absolute volume of talent. According to a Bank of Korea report, South Korea's specialized AI workforce was estimated at approximately 57,000 as of 2024. This falls far short of major advanced nations such as the United States (780,000) and the United Kingdom (110,000).
3. The Most Severe Problem: Talent is Leaving
Equally as serious as talent cultivation is the issue of "talent outflow." The structural problem of domestic AI talent leaking overseas is worsening.
South Korea's net inflow of AI talent is -0.36 per 10,000 residents, placing it among the lowest levels in the world, and domestic doctoral graduates planning overseas emigration have increased for three consecutive years, with 592 in 2023, 658 in 2024, and 709 in 2025.
The causes are complex, with compensation disparities acting as the core issue. As of 2024, workers with AI skills in South Korea earned about 6% higher wages than general workers. In contrast, the wage premium associated with possessing AI skills in the U.S. reached approximately 25%.
Uncertainty in career pathways also drives departures. Although the level of strategic science and technology—such as semiconductors, batteries, and AI—is high, the scarcity of career paths for talent to grow, combined with a lack of quality jobs, low compensation, and infringement on research autonomy, has effectively turned the country into a recruitment hunting ground for overseas markets.
For senior researchers, mandatory retirement structures are also problematic. While hitting retirement age in South Korea means individuals must completely stop working, testimony indicates that in China, even graduate students essential for science and engineering research are assigned to them. The issue of talent outflow in the science and technology sector was brought to light during the 2024 parliamentary audit when it was revealed that 149 KAIST professors simultaneously received emails regarding "China's Global Outstanding Scientist Invitation Program."
The government also takes this issue seriously. President Lee Jae-myung stated that "the issue of overseas talent outflow is a very serious matter for the state," and formulated a plan to select 20 national scientists annually for five years starting at the end of 2026, providing each with 100 million KRW in annual research activity support funds. However, experts point out that this level of support alone is unlikely to serve as a fundamental remedy.
4. Field Realities: Small and Medium Enterprises Remain an 'AI Blind Spot'
While government cultivation policies have been designed primarily around graduate schools and research institutions, an entirely different reality is unfolding in industrial fields.
According to an "SME AI Utilization Intention Survey" conducted by the Korea Federation of Small and Medium Business at the end of 2024, 94.7% of small and medium-sized enterprises (SMEs) do not adopt AI, and 80.7% of those cited that "AI is not necessary for our business." Responses indicating no future plans to adopt AI also reached 83.7%.
This is directly linked to the absence of AI talent. SMEs' utilization of AI often stops at external service contracts or one-off projects, establishing a structure devoid of internal personnel to operate and improve upon it. Experts emphasize that cultivating "AI managers" who understand internal corporate data and operations while taking charge of AI transformation is an essential prerequisite for securing national AI competitiveness.
The manufacturing field faces similar circumstances. Although domestic manufacturing is gradually adopting AI, the scope of utilization and establishment of corporate guidelines remain at an initial stage. Even after AI adoption, responses of "no change" remain dominant across digital transformation expenditures, labor costs, operating profits, and organizational changes, meaning substantive changes are still limited.
Ultimately, while AI talent cultivation policies remain concentrated at the top tier, actual industrial foundations continue a vicious cycle of failing to even adopt AI due to a lack of available personnel.
5. Structural Differences Compared to Advanced Nations
United States: Private Sector-Led, Ecosystem Self-Sufficiency
From 2013 to 2023, private AI investment in the U.S. reached 335.2 billion USD, overwhelming China (103.7 billion USD). The proportion of top 20% AI researchers operating in U.S. institutions also increased from 27% in 2019 to 38% in 2022. While the U.S. government provides direction, practical talent cultivation and ecosystem building are organically handled by big tech, startups, and universities. Massive private research labs such as OpenAI, Google DeepMind, and Meta AI independently attract and cultivate world-class talent.
China: Deploying State Strategy and Private Sector Simultaneously
From 2013 to 2024, the U.S. ranked first and China second in the number of new investment attraction deals for global AI companies, with the two countries accounting for roughly 70% of total investment volume. China's characteristic lies in the systematic linkage of state-led funding, industry-education system connections, and large-scale overseas talent attraction programs. China already accounts for approximately 70% of worldwide AI-related R&D patents.
UK and Singapore: Specialized Strategies
The UK maintains its position as the world's 3rd largest in private AI investment scale (approx. 4.5 billion USD as of 2024) while seeking differentiation through basic research-centric talent cultivation and regulatory leadership. Singapore ranks in the global top tier for AI talent concentration relative to population. Along with Israel, it stands out as a small nation that excels in talent acquisition and startup ecosystem formation.
South Korea's Position
The Korea Development Institute (KDI) diagnosed that while South Korea is concurrently pursuing various strategies—such as institutional streamlining, infrastructure expansion, and talent cultivation to leap into the world's top 3 AI powerhouses based on core technological capabilities like semiconductors and past ecosystem foundations—securing global technological leadership requires more sophisticated execution strategies, including strengthening private participation foundations, ensuring policy continuity, and industry-tailored AI utilization.
6. Areas Where South Korea Maintains Strengths
Critical diagnoses alone do not constitute a balanced perspective. Areas where South Korea holds practical strengths do exist.
First is ICT infrastructure.
World-class ultra-high-speed internet networks, semiconductor manufacturing capabilities, and data center infrastructure serve as critical foundations for AI research and application environments. South Korea's AI competitiveness receives high evaluations in infrastructure, development, government policy, and scale indices.
Second is the speed of technological catch-up.
The AI technology gap with the U.S. across fields stands at 1.3 years for learning intelligence (a 0.7-year reduction compared to 2018), 1.5 years for single intelligence (a 0.5-year reduction compared to 2018), and 1.0 year for complex intelligence (a 1.0-year reduction compared to 2018), making South Korea one of the countries with the largest shifts in technological gaps.
Third is the volume of Large Language Models (LLMs) possessed.
South Korea possesses 6 or more foundation models (LLMs) serving as the basis for generative AI, and AI startups account for over 60% of the total, raising expectations for an important role in future markets.
Fourth is the presence of world-class research institutions such as KAIST, Seoul National University, POSTECH, and GIST.
Institutions equipped with artificial intelligence and AI convergence graduate schools—such as Korea University, Pusan National University, Seoul National University, Sungkyunkwan University, Ajou University, POSTECH, Hanyang University, KAIST, and GIST—are leading AI education and research.
7. Structural Limitations: The Core of the Diagnosis
The structural limitations of South Korea's current AI talent cultivation environment can be condensed into four main points.