Chrome fixed more security flaws during the month of June than in the last two years thanks to AI

  Google Chrome has assured that it has corrected a total of 1,072 security flaws in the browser during the last month of June, exceeding the total number of bugs corrected in the last two years, thanks to the use of artificial intelligence (AI) tools.

Specifically, the technology giant has assured that in the last two updates, Chrome 149 and 150, released during the month of June, more security flaws have been fixed than the total number of bugs fixed in the 23 previous updates combined.

This was announced in a new report shared on its blog, in which Google details in depth its methods to make Chrome and the web “more secure in the age of AI”, precisely by using this technology to improve the detection, classification and application of patches of vulnerabilities.

“Large-scale language models (LLMs) are unlocking unprecedented capabilities for automated vulnerability discovery, far pushing the limits of human security expertise and demanding new approaches to stay ahead of attackers,” Google said.

In this sense, Google continues to deploy large-scale AI models to find and fix “hundreds of security flaws”, faster and with greater resilience to achieve “comprehensive remediation”.

Specifically, the Chrome security team has been using LLM since 2023, to increase the coverage and performance of security tests. Since 2025, they have also collaborated with DeepMind and Project Zero on tools such as Big Sleep, an AI-based vulnerability discovery agent that successfully found errors in the V8 JavaScript engine and graphics stack.

Likewise, earlier this year, Google recalled that they developed a vulnerability analysis system that used Gemini to detect security flaws in the entire Chrome source code with greater efficiency and fewer false positives.

Currently, Google has LLM that generate suitable solutions for most vulnerabilities, which increases the rate of security fixes in recent versions of Chrome, reaching 1,072 security flaws fixed in June.

In addition, the technology company has also highlighted that, in May alone, they prevented more than 20 vulnerabilities from reaching production, including a critical S1+ level problem.

Regarding the latter, it should be noted that it is due to the fact that AI tools for security are natively integrated into Google’s continuous integration (CI) system, which runs every 24 hours at all compilation levels to proactively detect security flaws.

Finally, the company has recalled that, once a correction to a bug is implemented, depending on its severity, they are integrated directly from the main repository to the active stable branch of Chrome, which is continually monitored to prevent new bugs and patch gaps.