LAMEHUG malware uses large language models for adaptive attacks.
LAMEHUG is a novel AI-driven malware that leverages large language models (LLMs) to execute dynamic reconnaissance and data exfiltration in real-time. It uses the Hugging Face API to generate tailored Windows command-shell instructions based on its environment. This capability complicates detection efforts for security operations centers. LAMEHUG is delivered via spear-phishing emails with malicious executables disguised as AI image-generation tools. Upon execution, it displays a prompt interface while running hidden threads to orchestrate malicious activities. A variant drops a decoy PDF to further deceive users.
Source: https://gbhackers.com/llm-driven-malware/ GBHackers