---
title: Mozilla: AI Can Find All Security Vulnerabilities in Firefox
url: https://www.elseif.net/mozilla-ai-can-find-all-security-vulnerabilities-in-firefox
published: 2026-11-06T16:04:48+00:00
language: en
section: Models
source: https://www.security.nl/posting/960263/Mozilla%3A+AI+kan+alle+in+Firefox+aanwezige+beveiligingslekken+vinden?channel=rss
organizations: mozilla
publisher: elseif
---

# Mozilla: AI Can Find All Security Vulnerabilities in Firefox

Mozilla has utilized AI this year to release seven times more security updates for Firefox compared to the same period last year and expects that AI will ultimately identify all vulnerabilities in the browser. According to the developer of Firefox, AI models can now find vulnerabilities as effectively as the best security researchers, at a scale and speed that were not considered possible a year earlier.

This is promising news for defenders, as stated by Mozilla. The organization claims that the number of existing security vulnerabilities in Firefox is 'finite' and that increasingly powerful AI models are likely to soon detect all security issues in the browser. However, Mozilla warns of the risk of the 'patch gap.' This is the period when a solution for a vulnerability is made public, but the corresponding update has not yet been rolled out to users.

Mozilla notes that new versions must be tested and verified for stability before they become available to users. Additionally, updates are not installed immediately by all users. Based on a implemented fix, an attacker could determine where the problem lies and develop an exploit. To reduce the 'patch gap,' the Firefox developer now releases updates more frequently.

Another measure taken by Mozilla is to make regression tests public later. AI model GLM 5.3 was found to be much less effective with patch gaps when no regression tests are available to identify the underlying security issue, according to Mozilla. Therefore, regression tests are now kept confidential until security updates have been rolled out to users. Furthermore, the Firefox developer states that attackers do not yet have access to the AI models that Mozilla is using. The vulnerabilities they can find have largely already been discovered and resolved, according to the Firefox developer.