South Korea Ministry of Science and ICT Signs AI Safety and Cybersecurity MOU With Anthropic — Red-Team Evaluations of Autonomous Agents and Korean-Language Risk Assessment Planned Amid Ongoing Export Ban

South Korea's Ministry of Science and ICT signed an AI safety and cybersecurity MOU with Anthropic on June 18, directing cooperation toward red-team evaluations of autonomous AI agents, Korean-language model safety and misuse risk assessment, and joint analysis of AI's impact on cyber offense and defense operations. The agreement deepens institutional ties with South Korea's Korea AI Safety Institute, building on February discussions between South Korea's Deputy Prime Minister and Anthropic CEO at the AI Action Summit in India. The partnership carries an immediate complication: the US export-control directive blocking foreign nationals from Anthropic's frontier models since June 12 directly threatens the Korean-language safety evaluations the MOU envisions.