AI Enters the 'Semiconductor Security War'... Samsung, SK hynix Deploy Mythos

South Korea Joins Project Glasswing for U.S. Allied Nations
Moves to Prevent Leakage of Semiconductor Design and Process Information

South Korea’s key semiconductor companies, including Samsung Electronics and SK hynix, have begun utilizing "Claude Mythos," an ultra-high-performance cybersecurity AI developed by U.S. artificial intelligence (AI) company Anthropic, to assess potential security threats. There is growing attention on whether this will mark a turning point in establishing a new cyber defense system aimed at preventing information leaks in the country’s critical industries.

According to the Ministry of Science and ICT on June 3, the Korea Internet & Security Agency (KISA) recently participated in "Project Glasswing," operated by Anthropic, and has secured access to Mythos. On June 2 (local time), Anthropic also announced plans to expand "Project Glasswing" to include companies and organizations from over 15 U.S. allied countries. "Project Glasswing" is operated to identify security vulnerabilities using Mythos.

The Ministry of Science and ICT explained that it has maintained ongoing collaboration with Anthropic, and through KISA’s participation in the project, has established a foundation for major South Korean companies and organizations to leverage cutting-edge AI-based security inspection systems.

In particular, it was confirmed that leading domestic semiconductor companies, such as Samsung Electronics and SK hynix, are among the participants. The intention is interpreted as proactively identifying and blocking security weaknesses that could lead to leaks of advanced semiconductor design and manufacturing process information by utilizing Mythos.

Mythos is Anthropic’s next-generation AI model, which was released earlier this year. It is characterized by its specialization in detecting security vulnerabilities and analyzing code. Immediately after its release, Mythos shocked the IT and cybersecurity industries by uncovering vulnerabilities that had gone undetected for extended periods in major operating systems and browsers. Due to its extremely powerful capabilities, its public release has been restricted.

In response, Anthropic launched Project Glasswing in April, centered around the U.S. government and security companies. Since then, Mythos has been used to discover over 10,000 critical software vulnerabilities.

Industry experts point out that, given that advanced semiconductor design information and manufacturing process technology are considered national strategic assets, the leakage of such information could impact not only individual companies but also national economic security and industrial safety.

Kim Joungho, professor of electrical and electronic engineering at KAIST, explained, "AI is now serving as both the sword and shield in cybersecurity. The use of AI is essential to prevent the leakage of critical semiconductor technologies." Professor Kim also predicted, "Since Anthropic itself requires advanced memory semiconductors from Samsung Electronics and SK hynix to build its AI infrastructure, this collaboration could serve as an opportunity to strengthen ties between the two sides."

This content was produced with the assistance of AI translation services.