How to Use Shadow AI

Prev Next

Shadow AI refers to the use of generative AI tools that are not formally approved, governed, or visible to security teams.

As employees adopt new AI tools to improve productivity, usage often happens outside of sanctioned platforms. This creates blind spots where sensitive data may be shared without oversight or enforcement.

The Shadow AI page in Verax provides visibility into these tools and allows you to control their usage.

Note
Shadow AI discovery requires AI traffic to be routed through Verax. Tools that do not pass through Verax cannot be detected or controlled.


What Is Shadow AI?

Shadow AI includes any generative AI tool that:

  • Is used by employees without formal approval

  • Is not explicitly governed by organizational policies

  • Falls outside of existing security controls

Examples include:

  • Public AI chat tools

  • Browser-based AI assistants

  • Embedded AI features in third-party applications


How Verax Detects Shadow AI

Verax monitors AI-related traffic routed through the platform and identifies generative AI tools based on observed usage patterns.

When Verax detects usage of a generative AI tool, it records and displays that tool on the Shadow AI page, even if the tool has not been explicitly integrated or approved.

This allows security teams to discover AI usage that would otherwise remain hidden.


Understanding the Shadow AI Page

The Shadow AI page lists generative AI tools detected in your environment.

For each tool, the following information is displayed:

  • Name
    The name of the generative AI tool.

  • Category
    The type of tool, such as Native AI, Code Assistant, or Platform.

  • Last Used
    The most recent time the tool was observed in use.

  • Allow
    Indicates whether the tool is currently allowed or blocked.


Allowing and Blocking Tools

By default:

  • Generative AI tools explicitly integrated with Verax are allowed

  • All other detected tools are blocked

You can control access using the Allow toggle:

  • Allow enabled
    The tool is permitted for use.

  • Allow disabled
    The tool is blocked.

Blocking a tool prevents users from interacting with it through Verax.


Relationship to Rules

The Shadow AI page provides tool-level visibility and control.

For more granular enforcement, Verax rules allow you to:

  • Apply access controls based on users or groups

  • Restrict usage by tool category

  • Enforce conditions such as sensitivity or topic similarity

Allowing a tool on the Shadow AI page does not bypass rule enforcement. All allowed tools are still evaluated against your configured rules.


Security Considerations

Allowing a generative AI tool that is not explicitly governed may introduce risk.

Important considerations:

  • Newly discovered tools may not enforce the same data protections

  • Content sent to these tools may not be sanitized or controlled

  • Allowing a tool expands the AI attack surface

Security teams should review newly detected tools before allowing them.


Recommended Workflow

A common workflow for managing Shadow AI is:

  1. Review newly detected tools

  2. Identify tools that require governance

  3. Block tools that present unacceptable risk

  4. Allow approved tools

  5. Create rules to enforce how approved tools can be used


Summary

The Shadow AI page helps you identify and control generative AI tools that operate outside traditional security boundaries.

By combining discovery with rule-based enforcement, Verax enables organizations to reduce Shadow AI risk while maintaining visibility and control over AI usage.