Resource Type: Azure > AI Foundry > RAI Policy

An AI Foundry RAI (Responsible AI) Policy defines content filtering and moderation rules for an AI Foundry Account. RAI Policies apply to deployments that reference them and enforce filters for categories such as Hate, Sexual, Violence, Self-harm, Jailbreak, Protected Material (Text/Code), and Profanity, with configurable severity thresholds (Low/Medium/High), blocking behavior, and source (Prompt/Completion). Azure auto-creates Microsoft.Default and Microsoft.DefaultV2 system-managed policies on each account; customer-managed policies are user-defined. This mod follows the Microsoft.CognitiveServices/accounts/raiPolicies ARM resource type name.

Resource Context

RAI Policy is a part of the AI Foundry service.

Each RAI Policy lives under an Account.

Controls

The primary controls for Azure > AI Foundry > RAI Policy are:

Quick Actions

In Your Workspace

Developers

Resource Type URI

tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy

Category URI

tmod:@turbot/turbot#/resource/categories/security

GraphQL

query resource(id: "tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy") { … }

query resourceActivities(filter: "resourceId:'tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy'") { … }

mutation createResource(input: { … })

mutation updateResource(input: { … })

CLI

Get Resource

turbot graphql resource --id "tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy"

Steampipe Query

Get Resource

select * from guardrails_resource where resource_type_uri = 'tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy';

Get Policy Settings (By Resource ID)

select * from guardrails_policy_setting where filter = 'resourceTypeId:"tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy"';

Get Resource Notification

select * from guardrails_notification where resource_type_uri = 'tmod:@turbot/azure-aifoundry#/resource/types/raiPolicy' and notification_type in ('resource_updated', 'resource_created');