Google Uses AI to Accelerate Chrome Security Patching

Google is overhauling the security infrastructure of the Chrome browser by integrating large language models, including Gemini, to detect and patch vulnerabilities. This AI-driven approach has triggered a record surge in bug fixes; Chrome versions 149 and 150 contained 1,072 fixes combined, a total that exceeds the previous 23 major releases.

To manage this increased volume of updates and mitigate the risk of N-day attacks, the company is accelerating its release cycle to every two weeks starting in September and is piloting security updates twice per week. A core part of this strategy is the development of dynamic patching. This mechanism utilizes Chrome's multi-process architecture to replace background child processes, such as the Renderer and GPU, with updated binaries on the fly, eliminating the need for a full browser restart in most cases.

While the full dynamic patching system is still in development, Google has launched a version for macOS in Chrome 150 that allows automatic restarts during windowless states. These changes aim to counter AI-powered threats that have transformed vulnerability discovery into an automated, industrial-scale operation.