The North Korean hacking group Kimsuky has developed local AI tools to enhance its cyberattack capabilities, as reported by Genians Security Center on August 10, 2026. The group is utilizing large language models (LLMs)…
Microsoft researchers revealed that a benign prompt can disable safety features in 15 major language models. The technique, known as GRP-Obliteration, exploits a training method called Group Relative Policy…