New AI policy chief warns of export control risks, vows push for homegrown AI

Ha Jung-woo, the new standing vice chairman of the Presidential Council on National Artificial Intelligence Strategy (NAIS), said Monday that Korea must continue developing its own AI capabilities to mitigate risks such as potential export controls on top-tier AI models, citing a U.S. export control on Anthropic's Mythos 5 model in June. He emphasized that disparities in access to leading AI models could cause gaps in national capabilities, and that Korea needs to maintain its own capabilities to make frontier AI models, whether through the government's national foundation model project or other initiatives.

Ha Jung-woo, the new standing vice chairman of the Presidential Council on National Artificial Intelligence Strategy NAIS , said Monday that Korea must continue efforts to develop competitive AI capabilities of its own to mitigate future risks such as potential export controls on top-tier artificial intelligence AI models. During a press conference marking his appointment, Ha said export controls on top-tier frontier AI models could appear again at any time, citing a U.S. export control on Anthropic’s Mythos 5 model in June that spread fears that advanced AI tools have become strategic assets. “Disparities in access to leading AI models could cause gaps in national capabilities,” he said. “Whether it is the government’s ongoing national foundation model project or other types of initiatives, Korea needs to maintain its own capabilities to make frontier AI models.” Ha’s comments came as the national foundation model project, an initiative to foster domestic frontier-level AI models using locally developed technology, faces growing questions over whether the models are be