North Korean Hacker Group Builds Local LLM Environment, AI Attack Capability Enhanced

August 10th - Network security company Genians disclosed on August 10th that North Korea's Reconnaissance General Bureau-affiliated hacker group Kimsuky has made a significant advancement in Generative AI applications. They have not only used AI for phishing bait generation but have also further developed a standalone on-premises Large Language Model (LLM) environment and a Retrieval-Augmented Generation (RAG) system. Specific tools include Ollama, GPT4All, Msty, and other local LLM management tools, along with the AI code editor Cursor, with the intention of achieving intelligence extraction and automated attack processes. (Yonhap News Agency)