Agentic AI: How It Scales Cybercrime & Evades Defenses

Artificial intelligence is now giving cybercriminals the ability to run operations with less human oversight. John Hultquist, chief analyst at Google Threat Intelligence Group, states that agentic AI provides speed, scale, and autonomy for these activities. Here's the thing: threat actors are using AI beyond just creating phishing emails. They're now deploying AI models during intrusions to navigate networks, select attack paths, and automate technical tasks. This agentic AI can perform actions independently, allowing criminals to expand their operations using stolen or low-cost access to AI services. What's interesting is this shift challenges traditional signature-based defenses. Attackers can use AI to rewrite malware and alter recognizable elements, reducing the value of indicators from repeated behavior. Security teams will need to focus more on behavioral detection and use AI to spot unusual network activity. The bottom line is that defenders also need to embrace AI for behavioral detection, vulnerability discovery, and malware analysis to counter these evolving threats.