Fortinet has expanded AI security functions for its FortiGate firewalls. SiliconANGLE reported on Oct. 26 that Fortinet launched the "FortiGuard Inline AI Security Service" and also strengthened the functions of its AI gateway, FortiAIgate.

The new service is provided to existing FortiGate firewalls through the Fortinet Cloud. Existing customers can check in real time prompts heading to large language models, agents and model context protocol tools without building an additional AI security stack.

FortiGate already could detect shadow AI use and apply data-loss-prevention rules to AI traffic. This time it added a function that inspects the interaction content itself. Sensitive information in transit is masked.

FortiAIgate is a gateway that sits between applications and AI models. Fortinet said, "Existing security tools that only look at IP addresses and signatures cannot stop prompts that hide instructions targeting agents." It said, "FortiAIgate evaluates each request in real time by also considering the requester's role. It inspects images, video and code as well as text, and blocks 35 types of personal and confidential information before they leave the organization."

Administrators can set cost limits by team or by user. Fortinet said it blocks prompts designed to consume many tokens before token use, and can also route requests to an appropriate model based on content, policy and cost criteria.

Fortinet acquired AI security startup VirtuAI in August and has already reflected VirtuAI technology in FortiAIgate guardrails.