Researchers Show Copilot and Grok Can Be Abused as Malware C2 Proxies

AI-enabled web-browsing assistants can be abused as stealthy command-and-control relays that blend attacker communications into legitimate enterprise traffic. Check Point demonstrated this “AI as a C2 proxy” technique against Microsoft Copilot and xAI Grok, warning it can enable AI-assisted malware operations and dynamic, evasive implants. #MicrosoftCopilot #xAIGrok

Keypoints

AI browsing and URL-fetch features can be turned into stealthy C2 channels that tunnel commands and data.
Check Point labeled the method “AI as a C2 proxy” after demonstrating it against Microsoft Copilot and xAI Grok.
The attack uses anonymous web access and crafted prompts to fetch attacker-controlled content without API keys or accounts.
An initial host compromise is required, with malware using the AI agent to receive commands and exfiltrate information.
Palo Alto Unit 42 research shows related LLM-based techniques can generate malicious JavaScript and enable Last Mile Reassembly-style phishing and evasion.