December 1, 2025—Announcing the official release of AI Guard, an application programming
interface (API) endpoint that provides real-time protection by monitoring and filtering
AI system inputs and outputs. This solution addresses the unique security challenges
that arise when deploying AI at scale.
Key features include:
-
Guardrails for inbound requests to block malicious input and prevent misuse.
-
Scanning for harmful content generation and sensitive information leakage.
-
Protection against AI model poisoning.
-
Sensitive data protection and privacy compliance.
-
Infrastructure security to manage risks in applications powered by LLMs, preventing abuse and misuse.
AI Guard enables users to safely implement AI capabilities while maintaining security
standards and compliance requirements. For more information, see AI Application Security.