Fortinet releases security hardening updates for multiple products

two silver swords on shield — Patrick Hendry / Unsplash

Fortinet released a security hardening update for FortiOS, FortiManager, FortiAnalyzer, FortiMail, and FortiAuthenticator. This update addresses critical and high vulnerabilities discovered with the help of artificial intelligence. Fortinet will not disclose details of the vulnerabilities until November 10, 2026. The vulnerabilities were identified using an AI model and confirmed by Fortinet's team. However, attackers can also access AI tools to find and exploit vulnerabilities. Fortinet expects the time between vulnerability disclosure and actual attacks to shorten. Once updates are released, patches reveal which code sections were fixed, making unpatched devices easier targets. Fortinet products are frequent targets for attackers. Some vulnerabilities may allow remote attacks without authentication.

Fortinet warns that AI can link low-severity vulnerabilities, making them more dangerous. The updated versions are available for FortiOS, FortiManager, and FortiAnalyzer 8.0, 7.6, 7.4, and 7.2. FortiMail 8.0, 7.6, and 7.4 are also updated, with FortiMail 7.2 users advised to upgrade to 7.4 or higher. FortiAuthenticator 8.0 and 6.6 are updated as well. Fortinet urges administrators to upgrade immediately, prioritizing internet-connected systems. Temporary risk reduction measures include limiting admin access to trusted IP addresses, avoiding internet exposure of FortiManager, using strong unique credentials, and enforcing multi-factor authentication for admin accounts. These measures do not replace the need for upgrades. This update does not address the PHP2Shell vulnerability, as none of the five products are affected.

Fortinet's FortiGuard PSIRT team used frontier AI models and on-premises AI for security testing over several months. They tested without source code, analyzing firmware to map operations and potential attack points. Penetration testing and red teaming confirmed the findings. AI effectively compared different firmware versions and created proof-of-concept exploits for fixed vulnerabilities. Most findings came from AI analyzing Fortinet's source code, which attackers typically cannot access. Over 100 tests across 30 product families showed 40 percent were real vulnerabilities, 30 percent were false positives, and 30 percent were security improvement points, not exploitable vulnerabilities.