Permissive AI access and limited monitoring could allow malware to hide within trusted enterprise traffic, thereby accelerating adaptive, AI-driven attacks.