Minjae Seo

Minjae Seo

Minjae Seo is a Research Scientist at the ETRI Korea AI Safety Institute (Korea AISI), working on AI safety, network security, and the adversarial robustness of frontier AI models. His earlier research focused on privacy protection in Tor, blockchain, and distributed systems. Broadening his scope to frontier AI models, he proposed the first code-patching agent addressing knowledge-cutoff vulnerabilities unseen by code-generation models, and introduced the first political jailbreaking agent for frontier text-to-image (T2I) systems. At Korea AISI, he contributes to frontier research on safe and trustworthy AI, focusing on model vulnerabilities, privacy-preserving systems, and generative AI security.

Share this speaker
Share This Speaker
  • Organization
    Korea AI Safety Institute (Korea AISI)
  • Profession
    Research Scientist
Related sessions
In person
7 July 2026
14:00 - 17:15
EST - New York
CST - Beijing
PST - Los Angeles
AWST - Perth, Australia

Are you sure you want to remove this speaker?