A new malware named LameHug has been identified as the first AI-powered malware linked to Russia’s APT28 cyber espionage group. Discovered by Ukraine CERT, LameHug uses large language models (LLMs), specifically from Alibaba and accessed through Hugging Face, to generate and execute Windows system commands in real time without relying on pre-written scripts. The malware is delivered via ZIP files and targets Windows machines to steal data. Security researchers characterize LameHug as a live test run reflecting ongoing research and development in AI-driven cyberattacks. Experts note that generative AI is accelerating attack lifecycles from days to minutes, raising concerns about the evolving sophistication of cyber threats. Carnegie Mellon University researchers have also demonstrated that LLMs can autonomously plan and execute complex cyberattacks without human intervention, underscoring the challenges in defending against AI-enabled threats.