Google found attackers using AI to automatically analyze patches and execute exploits.
TechFlow reports that on November 16, according to The Defiant, Google's threat intelligence team discovered that attackers are using an AI framework named exploit_agent to analyze security patches, write, and execute exploit code without human intervention.
Researchers discovered the framework along with data stolen from victims on an exposed relay server. The framework compares code before and after the release of web application patches, identifies potential vulnerabilities, and automatically executes exploits; the server also hosted scripts for internal network reconnaissance, brute-forcing, and installing webshells.
Google stated that victim telemetry data indicates the operators were interested in websites in the French media, entertainment, and education sectors. The company has not attributed the framework to any specific group.





