Google's Threat Intelligence Group (GTIG) has discovered an AI-powered system capable of identifying and exploiting software vulnerabilities without human intervention.
Google's Threat Intelligence Group (GTIG) has discovered an AI-powered system capable of identifying and exploiting software vulnerabilities without human intervention.
Dubbed exploit_agent, the framework compares code changes across software updates, identifies potential weaknesses, generates exploits, and executes them automatically. Google says this marks its first observation of threat actors using fully automated AI workflows for vulnerability discovery and exploitation.