In record time, artificial intelligence reveals 14 critical security vulnerabilities in the Firefox browser.
Artificial intelligence can be a valuable ally in cybersecurity. In this regard, the organization that created the Mozilla Firefox browser revealed that an AI system was able to detect over 100 vulnerabilities in the browser in just two weeks, including 14 critical ones.
This experiment is part of a collaboration between Mozilla and Anthropic's Frontier Red team, which used advanced models such as Claude AI to analyze the browser's code for errors.
According to Mozilla, the AI company recently contacted them after achieving promising results using a new AI-assisted bug-finding method.
The analysis focused primarily on the browser's JavaScript engine. This component is responsible for executing a large portion of the code used in modern web pages, so any security vulnerability in it could become an entry point for cyberattacks.
Furthermore, Firefox being an open-source project makes it an ideal testing environment for new analysis techniques.
During this process, the AI system was able to detect several security vulnerabilities and create miniature test cases that allowed developers to quickly reproduce errors.
14 critical vulnerabilities were confirmed, resulting in 22 different vulnerability tracer identifiers (CVEs).
In any case, there is no need to worry, as Mozilla confirms that all these problems have already been fixed in the latest version of the browser.
Mozilla emphasized that this approach differs significantly from other attempts to use artificial intelligence for bug detection. Some open-source projects have been forced to restrict or even ban AI-generated reports due to the large number of low-quality reports submitted by users seeking easy rewards in bug bounty programs.
In this case, the AI model not only detected flaws that are usually detected by automated techniques, but also logical errors that are difficult to identify by traditional methods.