Policy Configuration
Define, configure, and enforce responsible AI policies within your projects
Overview
The Policy Configuration module enables users and administrators to define and manage responsible AI policies across projects. This ensures alignment with compliance, security, and ethical standards using AltrumAI’s flexible and intuitive configuration system.
Accessing Policy Configuration
- From the Dashboard, select a project to open the Project Overview page.
- From here, you can either:
- Edit an existing policy from the Policies Table, or
- Click the Edit Policies button to begin configuring new policies.
This will take you to the Policy Configuration screen, which contains three tabs:
- Configure Policies
- Configure Response
- Test Policies
Configuring a Policy
Step 1: Navigate to the Configure Policies Tab
Within the Configure Policies tab, you’ll see a list of policies provided by AltrumAI, such as:
- Data Privacy
- Bias & Fairness
- Toxicity
- Hallucinations
- Competitor Mention
- and more
Step 2: Select a Policy
- Click on a policy from the list on the left.
- The selected policy’s configuration options will appear on the right-hand panel.
Step 3: Configure the Policy
Each policy has its own set of configuration options. Common configuration steps include:
- Enable the Policy using the toggle at the top of the panel.
- Choose an Enforcement Mode:
- Log Only – Log violations without blocking interactions.
- Log and Override – Block content that violates the policy.
- Adjust Sensitivity Thresholds (if applicable).
- Select relevant Topics, Entities, or Regulations (depending on the policy type).
For detailed instructions on each policy type, refer to the Policy Overview section of the documentation.
Step 4: Save the Configuration
- Click Save Changes to store your configuration.
- Your updates will now be associated with the project, but not yet active until applied.
Testing Policies
Once you’ve configured one or more policies, navigate to the Test Policies tab.
This is a sandbox environment where you can:
- Interact with a test chatbot.
- See your configured policies in action.
- Observe how prompts and responses are flagged, modified, or blocked in real time.
This is a valuable step to validate policy behavior before applying changes to a live project.
Applying Configured Policies
When you’re satisfied with your configuration and testing:
- Navigate to the Integration page.
- If needed, retrieve your Project API Key from the integration settings.
- Review the Summary page where you can:
- View all configured policies
- Make additional changes if required
- Update existing configurations
- Once you’re happy with all changes, click View Project to return to the Project Dashboard.
Editing or Disabling Policies
To update or turn off a policy:
- From the Project Dashboard, click Edit Policies to access the Configure Policies page.
- For each policy:
- Toggle the Enable button On or Off based on your required configuration
- Make any additional configuration changes as needed
- Click Save Changes to confirm your updates.
Best Practices
- Start with Log Only mode to monitor behavior without disruption.
- Use the Test Policies tab to validate before applying changes.
- Regularly review violation logs and refine configurations as needed.
- Apply stricter enforcement to regulated or high-risk projects.