South Korea boasts world-class artificial intelligence (AI) technology infrastructure but lags significantly in securing core AI talent, according to a new report. While the government has expanded AI graduate schools and increased research and development (R&D) investment for years, critics say the talent development approach itself is flawed, citing talent outflow, a supply-demand "mismatch" between industry and universities, and the concentration of top talent in medical schools.
According to the report "Strategies and Policy Tasks for Nurturing AI Talent," released Monday by the National Assembly Futures Institute, Korea ranked fifth in overall AI competitiveness under the Global AI Index but placed just 13th in the talent category. In particular, according to the "2026 AI Index Report" by Stanford University's Institute for Human-Centered AI (HAI), Korea's net AI talent inflow rate stood at minus 0.36 per 10,000 people, classifying it as a net outflow country where talent leaves for abroad, unlike major advanced nations.
The report defined Korea's AI talent problem not as a simple labor shortage but as a "structural failure." Since announcing its "AI National Strategy" in 2019, the government has pursued policies such as expanding AI graduate schools, increasing enrollment in advanced disciplines, and broadening military service exemptions, but cases of professors and doctoral-level researchers at major domestic universities moving abroad are rising.
In this regard, the report points out that the AI talent supply structure is excessively constrained by "balanced regional development" policy. In fact, universities in the capital region are subject to enrollment caps under the Seoul Metropolitan Area Readjustment Planning Act, enacted in 1982, making it impossible to adjust the quotas of AI-related departments to match industry demand. On top of this, most universities structure professor evaluation criteria around paper-focused "quantitative metrics," making it difficult for professionals working in the AI industry to teach.
The qualitative limitations of AI education are also raised. With standards for AI-related curricula not yet established, teaching varies widely across universities. Moreover, application-centered curricula are mainly operated with the goal of early cultivation of personnel ready for industry, so programs that balance basic, advanced technology, ethics, and application fields are rare. The tendency toward theory-centered education due to a shortage of AI specialist professors is also cited as a problem.
The structural disconnect in industry-academia cooperation is also a key issue. Major companies in AI fields are filling related personnel needs by directly recruiting talent or running their own training programs, judging that university capabilities do not meet industry requirements. Small and medium-sized enterprises, citing the time and cost burdens of joint industry-academia R&D, face clear limits in running such programs and are reluctant to participate actively. Meanwhile, students prefer participating in industry-academia cooperation programs led by large companies and are reluctant to join such programs with small or venture firms, creating a mismatch between market demand and supply. In addition, professors are not actively engaging in industry-academia cooperation due to a paper-focused performance evaluation system, raising the issue of poor learning content stemming from a breakdown in information exchange between universities and industry.
The absence of a control tower for AI talent is also a problem. AI talent development projects are pursued separately by central ministries such as the Ministry of Science and ICT, the Ministry of Employment and Labor, the Ministry of Education, and the Ministry of Trade, Industry and Energy, lowering policy efficiency. The current approach, in which the government induces competition among universities through large-scale financial support, has a problem in that it leads universities to design related programs according to funding criteria rather than specialized policies for cultivating AI talent, potentially producing uniform, cookie-cutter AI personnel.
The report also points out that AI talent policy focuses excessively on the scale of talent, such as "how many were produced." Even when AI specialists are cultivated, the effectiveness of related policies is halved if they do not settle domestically, yet efforts are concentrated solely on personnel cultivation in higher education, the analysis found. In fact, the United States attracts global talent through world-class research environments and compensation systems, while China supports returning researchers with packages covering research funds, housing costs, and children's education.
The report particularly noted China's case. Since its "Next Generation AI Development Plan" in 2017, China has maintained a consistent strategy for more than 10 years, expanding AI-major universities from 35 to over 600. It has also established institution-level "block funding" and multi-year research funding systems so that researchers do not cling to short-term project contracts, and it evaluates research achievements based on technology transfer and industrial application effects rather than the number of papers.
The report stated, "For advanced AI fields alone, capital region enrollment caps should be eased, and excellent professors should be actively secured through measures such as introducing a national chair professor system," adding, "The research environment itself must also be improved by building national-level AI computing infrastructure and a public data opening system." It continued, "There is a need to designate regional AI research hubs and consider introducing a block funding system that provides researchers with long-term, stable research funds," noting, "Return support packages should be provided to Korean researchers working abroad, and AI-specialized visas and tax benefits should be expanded for foreign researchers."
The report also stated, "The Ministry of Science and ICT (319.4 billion won) and the Ministry of Education (334.8 billion won) independently execute AI talent development budgets for similar targets, and as multiple ministries pursue their own projects separately, a structure with much overlap and insufficient synergy has become entrenched," adding, "Through measures such as revising the AI Basic Act, the National Artificial Intelligence Strategy Committee should function as the substantive control tower for AI talent policy." It continued, "It takes at least 8 to 10 years to produce a single high-level AI talent, but under the current three-to-five-year financial project structure, consistent investment that withstands this time is difficult to achieve," emphasizing, "Korea's manufacturing strengths, its semiconductor, robot, and automobile industry base, and its high educational enthusiasm remain valid assets, and a consistent strategy and structure must be created to connect these assets with the AI talent ecosystem."
