Fortinet Patches Nearly 100 Vulnerabilities, Crediting Frontier AI in Its Own Audit Pipeline
The 89 advisories issued on 10 November include roughly 17 critical flaws, with the vendor warning that AI has compressed the window from disclosure to exploit to days.
Fortinet released 89 security advisories on 10 November closing close to a hundred vulnerabilities, including roughly 17 critical and 80 high-severity issues, and said a significant share were found by a frontier AI model the company has integrated into its product security testing pipeline. The vendor said the model analyzes code paths, protocol handling and configuration logic at a depth and speed beyond conventional fuzzing and manual review, and thanked Anthropic, OpenAI, Google and SpaceX among external contributors.
The release spans the company's portfolio, with FortiMail advisories FG-IR-26-199 and FG-IR-26-200 covering path traversal and command injection flaws rated CVSS 9.8. Fortinet coordinated disclosure unusually: fixed builds were distributed to customers in October without detail, with specifics published only in November.
The company also issued a pointed warning on the offensive side: AI lets attackers diff patches into working exploits within hours to days rather than weeks, and it urged customers to treat this hardening release as a priority change. The pattern landed a day before Apache published a separate critical httpd fix credited to Anthropic's AI scanner.