Google Professional Cloud Security Engineer Question 365
Select 3Google Cloud PlatformYour organization is deploying a machine learning model on Google Cloud to analyze user financial data. To ensure the security and privacy of the data and the model, you need to implement controls to prevent unintentional exploitation, such as data leakage or model inversion attacks. Which of the following actions should you take?
- A
Use Differential Privacy techniques to anonymize sensitive user data before training the model.
- B
Deploy the model on a public endpoint without requiring authentication for easier access by users.
- C
Apply encryption to both data at rest and data in transit during the model training and serving process.
- D
Implement AI model monitoring to detect and respond to unusual access patterns or adversarial inputs.
- E
Store training data in a publicly accessible storage bucket for faster data sharing between teams.
Show answer and explanation
Correct answers: A, C, D
Explanation
AI/ML systems must be secured to prevent unintended exploitation of data or models. Techniques like Differential Privacy protect sensitive data during training, while encryption secures data at all stages. Monitoring AI models ensures any abnormal behavior is detected and addressed promptly. Public endpoints or storage without authentication and access control introduce critical vulnerabilities, violating best practices for securing AI/ML systems.
- A. Correct.
Using Differential Privacy techniques is critical for anonymizing sensitive user data, reducing the risk of data leakage or re-identification attacks.
- B. Incorrect.
Deploying the model on a public endpoint without authentication introduces significant security risks, such as unauthorized access and exploitation of the model.
- C. Correct.
Encrypting data at rest and in transit ensures that sensitive information is protected from unauthorized access during storage and communication, mitigating potential security breaches.
- D. Correct.
AI model monitoring is essential to identify and mitigate potential threats like adversarial inputs or abnormal access patterns, enhancing the overall security of the system.
- E. Incorrect.
Storing training data in a publicly accessible storage bucket is a serious security risk, as it exposes sensitive data to unauthorized access and potential misuse.