AI is finding so many Chrome security flaws that Google may start updating it twice a week

Google is accelerating its Chrome update cycle due to AI tools uncovering a surge of security flaws in the browser's code. Versions 149 and 150, released back-to-back, included a combined 1,072 bug fixes, outpacing all previous updates combined. These AI-driven analyses can now scan code at unprecedented speeds, uncovering issues previously unnoticeable.

In one instance, a flaw hidden for 13 years was exposed, potentially enabling attackers to bypass Chrome's sandbox protections and access local files. This rapid discovery rate is prompting Google to reconsider its update approach. While the company has already reduced its release cycle to every two weeks, it is now testing a more frequent model, potentially updating Chrome twice a week.

This would minimize the window between identifying a vulnerability and deploying a fix. However, AI's rapid ability to locate vulnerabilities also poses a risk, as attackers could exploit weaknesses just as quickly. Faster patching can mitigate this risk, but more frequent updates also mean users may not install them promptly, leaving systems exposed longer than intended.

Google is addressing this issue with Chrome 150 on macOS, which can now update itself when no windows are open, eliminating the need for a manual restart in many cases. This "zero window restart" feature is part of Google's broader strategy to move toward continuous, seamless updates. The company is also exploring "dynamic patching," a technique that would allow updates to be applied without any restart at all, though no timeline has been set for when this capability might be available.

As AI's vulnerability detection capabilities improve, the pressure on Google to respond swiftly will likely intensify, potentially driving the browser towards a more continuous update model with minimal disruption to users.