South Korea Bets on “Sovereign AI” — but Finalists Used Chinese Code

South Korea’s drive to build a nationally independent artificial intelligence ecosystem has hit an unexpected hurdle. In the country’s flagship government AI competition, more than half of the finalists were found to be using foreign technologies, in some cases Chinese-origin code—a revelation that comes just as Seoul commits record funding to the effort. Parliament approved a ₩727.9 trillion ($495.8B) budget for 2026, and President Lee Jae-myung has more than tripled AI investment to ₩10.1 trillion ($6.9B). The government frames AI as a cornerstone of South Korea’s future economic competitiveness and national security.

A Costly Irony The controversy is striking because the competition’s explicit goal is to reduce dependence on U.S. and Chinese AI systems. Launched last June, the three-year program aims to select two domestic champions by 2027 to deliver models built solely on Korean-developed technologies. Reality has proven more complicated. Of the five finalists, three were found to have incorporated foreign code or components, including technology linked to China.

“Starting From Scratch Isn’t Practical” Companies involved argue that ignoring existing open-source ecosystems and rebuilding everything from zero is inefficient and unrealistic. Critics counter that relying on foreign tools creates security risks and undermines the very idea of “sovereign AI.” Harvard electrical engineering professor Gu-Yeon Wei, familiar with the competition, says an absolute ban on external code is unworkable: “If you give up open-source software, you lose an enormous amount of benefit.” Governments worldwide face the same dilemma: how to cut reliance on foreign tech without slowing innovation in a field that affects both economic power and defense.

Spotlight on Finalists: Upstage, Naver, and SK Telecom The sharpest scrutiny fell on Upstage, after Sionic AI CEO Ko Suk-hyun claimed seeable similarities to open-source systems from China’s Zhipu AI, including visible copyright notices in code. Upstage responded with a live verification session, presenting development logs to show its model was trained using in-house methods. The company acknowledged, however, that its inference code included open-source components commonly used worldwide. Ko later apologized. Attention then shifted to other finalists. Naver faced claims that its visual and audio encoders resembled products from Alibaba and OpenAI. SK Telecom was questioned over inference code similar to that of DeepSeek, another Chinese AI firm. Both Naver and SK Telecom admitted using standard external components, while stressing that their core learning and training engines were developed independently.

Rules Unclear, Debate Intensifies A central issue remains unresolved: the competition rules never clearly stated whether foreign open-source code is allowed. South Korea’s Ministry of Science has not issued new guidance since the controversy broke. Still, Science Minister Bae Kyung-hoon welcomed the debate: “Watching the technological discussions now energizing our AI industry, I see a bright future for Korean AI.”

What’s at Stake By 2027, the two winners must achieve at least 95% of the performance of leading global models from firms like OpenAI or Google. In return, they’ll receive government funding for data, talent, and access to critical AI chips. The episode underscores a broader truth: sovereign AI isn’t just about budgets. It’s a test of where to draw the line between openness and security, speed and control, and whether a truly national AI stack is feasible in a deeply globalized tech ecosystem.

#AI , #SouthKorea , #INNOVATION , #Geopolitics , #technews

Stay one step ahead – follow our profile and stay informed about everything important in the world of cryptocurrencies! Notice: ,The information and views presented in this article are intended solely for educational purposes and should not be taken as investment advice in any situation. The content of these pages should not be regarded as financial, investment, or any other form of advice. We caution that investing in cryptocurrencies can be risky and may lead to financial losses.“