New Framework Developed to Counter LLM-Driven Cyberattacks
/ 1 min read
New Defense Strategy Against LLM-Driven Cyberattacks Introduced. Researchers have developed Mantis, a defensive framework designed to counter cyberattacks powered by large language models (LLMs). Mantis utilizes adversarial inputs to disrupt attackers’ operations, either passively by confusing their LLMs or actively by compromising their machines. The framework has shown over 95% effectiveness in experiments and is available as an open-source tool to encourage further research and collaboration in cybersecurity.