Cybercriminals leverage autonomous AI agents for faster, larger attacks

Cybercriminals are increasingly utilizing artificial intelligence, moving beyond simple chatbot prompts to sophisticated autonomous, multi-agent frameworks capable of executing attacks with unprecedented speed and scale, as first reported by Channel Insider.

The Google Threat Intelligence Group (GTIG) reported that attackers are employing AI agents to automate and accelerate malicious activities. In one instance, a financially motivated attacker used an agent-driven operation to conduct a mass credential-harvesting campaign in under six hours, autonomously managing vulnerability scans, resolving bugs, and rotating IP addresses. This evolution poses significant risks to managed service providers (MSPs) and managed security service providers (MSSPs), as a single compromised privileged account or automation tool could expose multiple client environments rapidly.

Attackers are also targeting software supply chains, compromising open-source ecosystems to deploy malware like DUSTMAKER, which can evade AI coding agent trust checks. Furthermore, enterprise AI assets are becoming targets for extortion, while techniques like "LLMJacking" exploit cloud environments for unauthorized processing power. The speed of these AI-driven attacks outpaces traditional security operations center (SOC) response times, necessitating faster incident escalation procedures and enhanced credential isolation for channel partners.

Source: Channel Insider