Search Results: AIMalware

Google Threat Intelligence Group (GTIG) reveals a critical shift: threat actors are now deploying AI-powered malware that dynamically rewrites its own code mid-attack to evade detection. State-sponsored groups and cybercriminals are weaponizing LLMs like Gemini and Qwen for real-time obfuscation, command generation, and social engineering, signaling a dangerous evolution beyond simple productivity abuse.

The Nx 's1ngularity' supply chain attack leveraged AI-powered malware to hijack 2,180 GitHub accounts and expose 7,200 repositories. Attackers used LLM prompts to refine credential theft tactics in real-time, exploiting stolen tokens to escalate access. This incident underscores critical vulnerabilities in CI/CD pipelines and the dangerous evolution of AI-assisted cyber threats.

Ukraine's CERT-UA has exposed LameHug, a novel malware that leverages Alibaba Cloud's Qwen LLM via Hugging Face to dynamically craft data-theft commands on infected Windows systems. This innovation enables Russian state-backed group APT28 to adapt attacks in real-time, evading traditional defenses and signaling a dangerous evolution in cyber warfare.