AI Application Security

Views:

Scan your AI models for common attack techniques, attack objectives, and harmful or sensitive content in inputs and outputs.

Important

This is a "Pre-release" feature and is not considered an official release. Please review the Pre-release disclaimer before using the feature.

Note

This feature is not available in all regions.

AI Application Security enables you to intercept malicious inputs and block potentially harmful outputs from your AI models, which helps to prevent exploitative usage and maintain regulatory compliance. If your data needs to remain local, choose self-hosted instead of hosted by Trend Micro.

Go to Cloud Security → Security for AI Stack → AI Application Security and click Get started for the tool that meets your need:

AI Scanner: Scan your AI models for common attack techniques and objectives to prevent malicious use and ensure regulatory compliance.

For more information on configuring a scan, see Configure AI scan settings.
AI Guard: Configure settings to scan your AI usage for harmful content generation, sensitive information leakage, and prompt injections.

For more information on integrating AI Guard with your application, see Integrate AI Guard.