Google Vertex Integration for AI Governance Models

You can integrate an instance of Google's Vertex AI with OneTrust to enable automatic mapping of model attributes (name, description, and version) from Google Cloud Platform into corresponding fields in OneTrust. This offers Data Scientists a workstream that is more efficient than manual data entry.

Prerequisites

Roles

  • Eventarc Event Receiver (with eventarc.events.receiveEvent and eventarc.events.receiveAuditLogWritten)

  • Vertex AI Viewer

Permissions

Enable the following in Audit Logs:

  • Admin Read

  • Data Read

  • Data Write for Cloud AutoML API

APIs

  • Vertex AI API

  • Cloud AutoML API

  • Compute Engine API

  • Cloud Monitoring API

  • Cloud Logging API

  • Cloud Functions API

  • Cloud Run Admin API

  • Eventarc API

  • Cloud Build API

  • AI Platform Training & Prediction API

  • Serverless VPC Access API

  • And any other APIs that may be needed based on customer roles accessibility during the configuration and execution of the Cloud Function, if not already enabled.

To create a new project

  1. Go to the Google Cloud Platform console.
  2. Click the Create Project button.
  3. Enter a unique name for the new project.
  4. Click the Create button.

To set up cloud function

  1. In the Google Cloud Platform console, navigate to the Cloud Functions screen.
  2. Click the Create Function button.
  3. In the Basics section:

    • Select 2nd gen in the Environment field.

    • Provide a unique name for the function (e.g. trigger-onetrust-on-create-model)

    • Select associated region (e.g. us-central1 (Iowa)

  4. Add the Other Trigger type.
  5. In the Eventarc trigger section:

    • Select Google sources in the Trigger type field.

    • Select AI Platform (under Vertex AI) in the Event provider field.

    • Select either the UploadModel or CreateTrainingPipline in the Event field.

      CODE 
      google.cloud.aiplatform.ui.ModelService.UploadModel 
google.cloud.aiplatform.ui.PipelineService.CreateTrainingPipeline
  6. Choose the required Region and the newly created Service Account.
  7. Save the trigger.
  8. In the Runtime service account section, update the setting according to your requirements and select the new service account.
  9. For the environment variable, provide OT_URL as the name and the OneTrust Integration Workflow Webhook with the Authorization Token.
    CODE 
    https://example.app/integrationmanager/api/v1/webhook/02022c73-6b40-4b39-a20c-9c13ab5defa7?Authorization=Bearer ZjlhXXXjY2xxxTE5NxxxNjhkY2M2ZmE4YxxxN2M0NzxxxkZlRVRyMxxxUhFWHpBVxxxTzNjMjlKXXXGNDXxXnH=
    Note:

    For more information on Authorization Keys, see Managing OAuth 2.0 API Keys.

  10. In the Connections tab, set up the Egress settings to reference the newly created Serverless VPC Access Connector.
  11. Click the Next button.
  12. Update the source code in main.py and requirements.txt with the reference for creating models. To acquire the respective files for the import and create flows, please contact your account representative.
  13. Click the Deploy button to launch the cloud function.

To create the integration

  1. In the Integrations menu, select Gallery.

    The Integration Gallery screen appears.

  2. Use the search function to locate the Google Vertex tile.

    Integrations gallery screen
  3. Click the tile.

    The Gallery Details screen appears.

  4. Click the Add button for Google Vertex Model Registration.

    The Add Workflow Name modal appears.


    Google Vertex - Gallery details
  5. Complete the fields, then click the Create button. The Integration Builder screen appears.
    Note:

    You can select None for the Credential field.


    Add workflow name
  6. The Built-in Webhook Trigger action will look for the model details using the following sample body:
    JSON 
    {
    "id": "1264849589291188224@1",
    "name": "model-1",
    "description": "model-1 description",
    "version": "1"
}
  7. The POST Create an Object with Values Defined in the Request action will return the model's name, organization group ID, version, and description using our provided payload variables:

    POST Create an Object with Values Defined in the Request - Request tab
  8. Save and activate the integration. Whenever a model is now created and registered, a log will captured for the integration trigger and expected model name.

To create a new model in Vertex AI

  1. In the Google Cloud Platform console, navigate to the Vertex AI screen.
  2. Click the Model Registry tab.
  3. Click the Create Model button.
  4. Enter a unique name for the new model.
  5. Select a model type.
  6. Click the Create button.