Anthropic is set to meet with the South Korean government to discuss artificial intelligence safety and security issues, including its AI model “Mythos.”
Industry sources said May 6 that Michael Sellitto, Anthropic’s head of global policy, will visit South Korea next week and plans to meet with Bae Kyung-hoon, deputy prime minister and minister of science and ICT. A ministry official said, “We are coordinating a meeting next week between Deputy Prime Minister Bae and Anthropic.”
Sellitto is an AI policy and security specialist who previously served as cyber policy director at the U.S. White House National Security Council. At Anthropic, he oversees global policy and external relations.
Mythos is expected to be a key topic. The model highlights AI-based cybersecurity and vulnerability detection capabilities.
Anthropic on April 7 officially unveiled the Claude-based model, previously known by the codename “Capybara,” under the name Mythos. The company said that during seven weeks of testing, Mythos autonomously found more than 2,000 previously undisclosed software vulnerabilities, or zero-days, across major operating systems and browser environments.
In Mozilla’s Firefox browser, it identified 271 vulnerabilities. Of those, it was reported to have generated exploit code for 181.
Bae previously met with Anthropic CEO Dario Amodei in February at the “2026 India AI Impact Summit” held in India.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.