Exploitation of Generative AI for Covert Malware Command-and-Control (infosecurity-magazine.com)

• LLMs utilized as trusted relays for malware communication
• Evasion of traditional firewall and NDR systems
• High risk for enterprise and governmental infrastructure

"Recent research identifies a critical vulnerability where LLMs like Grok and Microsoft Copilot serve as communication relays for malicious software. By utilizing the legitimate infrastructure of these AI providers, attackers can bypass traditional network security perimeters. This technique masks C2 traffic as standard user-AI interaction, significantly complicating detection for security teams and threatening European digital sovereignty."