Databricks Generative AI Engineer Associate Question 242
Select 3You are deploying a generative AI model on Databricks to generate customer-specific insights. To ensure appropriate governance, you need to track and control access to the model's outputs and the underlying datasets. Which of the following steps should you implement to ensure proper governance on Databricks?
- A
Enable Unity Catalog to manage data permissions and lineage tracking.
- B
Use Databricks Workflows to automate model deployment without setting access controls.
- C
Set up role-based access controls (RBAC) to manage user and group permissions.
- D
Implement audit logging to monitor access and changes to datasets and model outputs.
- E
Allow unrestricted access to the model's outputs to encourage collaboration.
Show answer and explanation
Correct answers: A, C, D
Explanation
Proper governance on Databricks involves enabling Unity Catalog for centralized management, setting role-based access controls to manage permissions, and implementing audit logging to track access and modifications. These steps collectively ensure secure, compliant, and auditable usage of generative AI models and datasets. Unrestricted access or lack of access controls can result in security and compliance risks, and automation tools like Databricks Workflows require additional governance configurations.
- A. Correct.
Unity Catalog provides centralized governance, including fine-grained permissions and data lineage tracking, which is crucial for managing access and ensuring compliance.
- B. Incorrect.
While Databricks Workflows can help automate processes, it does not inherently set or enforce access controls, so additional steps are needed for governance.
- C. Correct.
Role-based access controls (RBAC) allow you to manage which users or groups have access to specific resources, ensuring secure and controlled access.
- D. Correct.
Audit logging is essential for monitoring access and changes, ensuring transparency and accountability in data and model usage.
- E. Incorrect.
Unrestricted access to outputs can lead to compliance issues, data misuse, and security risks, which goes against governance best practices.