Views:
AI Policy lets you create and manage policies that protect your AI applications. You can create multiple policies and assign each one to specific AI applications.
Each policy inspects AI prompts and responses as defined on the detection tabs:
  • Harmful content
  • Sensitive information
  • Prompt attacks
Each tab contains rules which apply one of the following actions to the corresponding content:
  • Allow & log: Allows content that matches the rule and records the detection in the logs.
  • Deny & log: Blocks content that matches the rule and records the detection in the logs.
For rules, detection levels, and categories in each tab, see AI policy detection settings.
You can do the following with AI Policy:
Action
Description
Create policy
Create a policy and configure its detection settings.
Assign policy
Assign a policy to one or more AI applications.
  1. On the AI Guard tab, select Manage AI Guard Policy.
  2. Click Add Assignment.
  3. Specify the alias name. Use a comma or new line to separate values. The policy applies when the AI application exactly matches at least one value.
  4. Select the AI policy to apply then click Save.
Related information