South Korea's government explored cooperation with Anthropic in artificial intelligence (AI) and cyber security. The government asked Anthropic to share related information before it discloses security vulnerabilities in major systems.

The Ministry of Science and ICT said on May 11 it met Michael Sellitto and other officials, including Anthropic's head of global policy, together with the foreign ministry, the National Intelligence Service, the Financial Services Commission, the AI Safety Institute, the Korea Internet & Security Agency (KISA) and the Financial Security Institute. They discussed securing AI safety and trust and steps to revitalise the domestic AI industry.

The meeting was a follow-up to AI cooperation discussed during talks between Vice Prime Minister Bae Kyung-hoon, who also serves as science and ICT minister, and Anthropic Chief Executive Dario Amodei at the 2026 India AI Impact Summit held in February.

With growing concern about threats that high-performance AI models such as Anthropic's "Mythos" could trigger, the two sides sought ways to cooperate in cyber security. They also put cooperation on South Korea's AI policies and systems on the table, including the AI Basic Act, to build a foundation for AI safety and trust.

The science ministry proposed cooperation with South Korean cyber security companies and institutions and asked for information sharing so South Korea can prepare in advance for vulnerability disclosures. The two sides plan to continue active cooperation and working-level discussions on the use of AI models in cyber security.

Ryu Je-myoung (류제명), the ministry's second vice minister, said, "We will actively cooperate with leading global AI companies to strengthen prevention and response systems to AI risks, including securing the safety of AI models and reinforcing cyber security capabilities."