Review AI security

AI Protection provides tools to visualize your organization's AI asset inventory, investigate security findings, and review proposed mitigations for risks and threats. This guide describes how to manage AI security using the AI Security dashboard and the Assets page.

Before you begin

• Configure AI Protection .
• Grant the required Identity and Access Management (IAM) roles to users who users who access the AI Securitydashboard and Assets > AI Resourcestab. For more information, see Required roles .

Use the AI Security dashboard

The AI Security dashboard lets you visualize your organization's AI asset inventory and review proposed mitigations for risks and threats.

Access the AI Security dashboard

To access the AI security dashboard, in the Google Cloud console, go to the Risk overview > AI securitypage:

Go to AI security

For more information, see AI Security dashboard .

Understand risk management for AI systems

This section identifies potential risks to your AI systems, displaying the top risks in your AI inventory.

To view a visualization and details for a specific issue, click the issue.

View AI threats

This section highlights the most recent threats to your AI resources.

In this section, you can do the following:

• To see all threats to your AI resources, click View all.
• To see detailed information about a specific threat, click the threat.

Visualize your AI inventory

This section provides a visual summary of your AI asset inventory, including projects that use generative AI, active first-party and third-party models, and the datasets used to train third-party models.

In this section, you can do the following:

• To view details about your inventory, click any node in the visualization.
• To view a detailed list of individual assets (such as Gemini models and custom-built models), click the tooltip.
• To view details about a model, click the model. This view displays details such as the endpoints where the model is hosted and the training dataset. If Sensitive Data Protection is enabled, the datasets view also indicates whether the dataset contains sensitive data.

Review AI framework findings summary

This section helps you assess findings generated by AI security policies and data security policies. It includes the following cards:

• Findings: This section displays a summary of findings generated by AI security policies and data security policies, such as Gemini model detected and Gemini model not protected by Model Armor .

  • To view all findings, click View all findings.
  • To view findings for a specific category, click the finding count next to that category.
  • To view details for an individual finding, click the finding.

• Sensitive data in Vertex AI datasets: Displays a summary of findings related to sensitive data in your datasets, as reported by Sensitive Data Protection. For more information, see Introduction to Gemini Enterprise Agent Platform .

Examine Model Armor findings

This section displays a graph of the total number of prompts and responses scanned by Model Armor along with the number of detected issues. The graph also provides summary statistics for specific issue types, such as prompt injection, jailbreak attempts, and sensitive data exposure.

For agentic workloads, you can configure Model Armor on Agent Gateway ( Preview ) to screen prompts and responses to and from agents.

This graph uses the metrics that Model Armor publishes to Cloud Monitoring. For more information, see Model Armor overview .

Inspect AI resources

In addition to the AI Security dashboard, you can inspect and query your organization's AI asset inventory on the Assetspage in the Google Cloud console by using the AI resourcestab.

The AI resourcestab provides a detailed, filterable view of your discovered AI systems and assets, including the following:

• Models
• Endpoints
• Data sources
• Pipelines
• Agents ( Preview )
• Model Context Protocol (MCP) servers ( Preview )
• Notebooks

You can use the AI resourcestab to assist with the following:

• Understand your AI inventory: View comprehensive lists of active AI assets across your organization, filtered by project, location, or resource type.
• Perform security investigations: Examine the full metadata, change history, and associated IAM policies of specific AI assets.
• Examine related findings: Quickly view all security findings associated with a specific AI asset to evaluate its risk exposure and determine necessary remediation steps.

For more information about viewing and filtering AI assets in the console, see Inspect assets that are monitored by Security Command Center .

What's next

• Learn more about AI Protection .
• Learn how to configure AI Protection .
• To assess risk, access dashboard data .