The government has announced that it is in talks with Anthropic and OpenAI to address cyber threats posed by the so-called Mythos Shock, an advanced artificial intelligence (AI) model developed by Anthropic.
Speaking to reporters at the 2026 World Information Technology (IT) Show in Seoul’s Convention & Exhibition Center (COEX) on Wednesday, Deputy Minister of Science and Information and Communications Technology (ICT) Ryu Je-myung stated that they’re actively pursuing official discussions with Anthropic and OpenAI regarding collaboration on security-focused AI models.
Anthropic’s Mythos and OpenAI’s GPT-5.4-Cyber have sent shockwaves through the global cybersecurity community. Unlike previous AI models that merely assisted with code reviews and vulnerability analysis, Mythos has reportedly evolved to design its own infiltration strategies and execute hacks autonomously, raising concerns that it could potentially rival the world’s top hackers.
Sources indicate that Anthropic and OpenAI have limited access to these models, sharing them only with select government agencies, security experts, and corporate clients.
Deputy Minister Ryu emphasized the need for direct information sharing with both companies, stating that even without full access to these models, it must remain vigilant about the potential impact of AI systems like Mythos on our cybersecurity landscape. It’s continuously monitoring developments in collaboration with industry leaders and Chief Information Security Officers (CISOs).
Later, Deputy Minister Ryu attended the World IT Show’s opening ceremony, touring exhibits from major tech players including Kakao, SK Telecom, KT, LG Uplus, Samsung Electronics, Samsung SDS, and Maum AI.
Expressing optimism about the industry’s direction, Ryu commented that it’s encouraging to see telecom companies making substantial investments in AI. Their proactive approach to developing data centers and building a robust AI full-stack ecosystem is particularly significant for the tech sector.
He concluded on a positive note, saying that it’s reassuring to see growing confidence in the domestic Neural Processing Units (NPUs). These technologies are poised to play a crucial role in advancing the AI ecosystem.