Google Professional Cloud Security Engineer Question 360
Select 4Google Cloud PlatformYour organization is deploying a machine learning model on Google Cloud that processes sensitive financial data. To comply with security best practices and regulations, which actions should you take to secure the AI workload?
- A
Enable Data Loss Prevention (DLP) API to detect and mask sensitive data before training the model.
- B
Use Vertex AI's built-in model monitoring to track and alert for anomalous predictions.
- C
Ensure the model is deployed with public access to enable external testing for vulnerabilities.
- D
Apply IAM roles with the principle of least privilege to control access to datasets and AI models.
- E
Encrypt training data and model artifacts using Cloud Key Management Service (KMS).
Show answer and explanation
Correct answers: A, B, D, E
Explanation
Securing AI workloads involves a combination of data protection, access control, and monitoring. Enabling DLP ensures sensitive data is protected during model training. Using Vertex AI's monitoring detects anomalies that could indicate potential threats. Applying IAM roles with the principle of least privilege restricts unauthorized access, and encryption ensures data confidentiality. Allowing public access to the model is a security risk and should be avoided.
- A. Correct.
Enabling the DLP API allows you to detect and mask sensitive data, ensuring that no personally identifiable or financial information is exposed during model training.
- B. Correct.
Vertex AI's model monitoring helps track anomalies, which can indicate potential security issues such as adversarial attacks or data drift.
- C. Incorrect.
Allowing public access to the model is a security risk and violates the principle of least privilege. Testing should be conducted in a controlled environment.
- D. Correct.
Applying IAM roles with the principle of least privilege ensures that only authorized personnel and services can access sensitive data and AI workloads.
- E. Correct.
Encrypting training data and model artifacts with Cloud KMS ensures data confidentiality and compliance with security regulations.