Google has fundamentally overhauled its security workflow by implementing AI-powered agents designed to identify and remediate vulnerabilities within the Chrome codebase. According to Security Affairs, the tech giant successfully addressed 1,072 security bugs across its last two browser releases—a figure that surpasses the total number of patches implemented during the previous 23 milestones combined. This shift signifies a major transition in how the company approaches software integrity and threat mitigation.

The deployment of these AI tools, which include custom agents built upon the Gemini model, has transformed the vulnerability management lifecycle. Beyond mere detection, the system now automates the triage process, including bug reproduction, duplicate filtering, and severity assignment. This automation saves developers hundreds of hours each month by streamlining the path from detection to deployment. Furthermore, Google has integrated specialized agents into its continuous integration system that generate, critique, and test patches before human engineers even perform a final review.

This aggressive move toward automation also aims to close the 'patch gap'—the dangerous window of time between a fix appearing in open-source code and its actual deployment to end-users. By integrating tools like BigSleep and CodeMender directly into their workflow, Google is proactively preventing potential exploits. These systems have already proven their value by blocking over 20 critical vulnerabilities in a single month. As AI continues to evolve, the company is refining its vulnerability reward programs to encourage external researchers to focus on complex, high-impact findings that supplement existing internal automated defenses.