Skip to content
VentureBeat·Indexed by this site · Originally published· Event →· Its place in the scan record →

Google says it found attackers had built an AI framework designed to automatically turn patches into exploits

Google: AI built to turn patches into exploits | VentureBeat

AI summary

Google's threat intelligence group disclosed on Monday that attackers had built an AI framework called exploit_agent that can automatically turn software patches into exploits. Google says this is the first time it has found attackers using an LLM framework to fully automate vulnerability discovery and exploitation without human involvement. The framework integrates multiple frontier and local LLMs and can continuously compare code before and after patches, identify potential vulnerabilities, and draft and execute exploits. It was found on an exposed pre-release server along with stolen victim data and post-compromise scripts, but the report does not confirm that the framework itself led to those compromises. Google did not identify the specific vulnerabilities exploited, the number of victims, or the measured time from patch to exploit, nor did it attribute the activity to a specific organization or country; Russian-language comments in configuration files describe it as a pilot, and victim telemetry shows the operators were focused on French media, entertainment, and education websites.

Read originalventurebeat.com