A retail company is developing a recommendation model. The data science team wants to use a public dataset that includes user purchase history. The legal team requires that the data be de-identified before it is used in training. The security team must ensure the de-identification process itself is secure. What is the most appropriate control?

Run the de-identification job in a dedicated VPC with no external IP addresses and use Cloud Data Loss Prevention (DLP) to inspect and transform the data.

Copy the dataset to a Compute Engine VM and run a custom Python script to remove personally identifiable information (PII) fields.

Encrypt the public dataset with a CMEK and store the decryption key in Cloud Key Management Service (Cloud KMS).

Grant the data science team Owner access to the Cloud Storage bucket containing the raw dataset so they can manage the de-identification process.

To protect data processed by a Gemini Enterprise Agent, which security control should be configured?

Data loss prevention (DLP) policies

Virtual private cloud (VPC) peering

Google CloudProfessional Cloud Security Engineer

Domain 3Objective 3

3.3 Securing AI Workloads PROFESSIONAL-CLOUD-SECURITY-ENGINEER Practice Questions (Page 4)

Part of the Ensuring data protection domain, which accounts for 23% of the PROFESSIONAL-CLOUD-SECURITY-ENGINEER exam.

33questions here

7free pages

9concepts

23%of the exam

Questions 16–20

16application · medium
A retail company is developing a recommendation model. The data science team wants to use a public dataset that includes user purchase history. The legal team requires that the data be de-identified before it is used in training. The security team must ensure the de-identification process itself is secure. What is the most appropriate control?
Select an answer first
17foundation · easy
A security engineer is reviewing threats to a machine learning system. Which attack involves an adversary submitting carefully crafted inputs during the training phase to cause the model to learn incorrect behaviors?
Select an answer first
18foundation · easy
In a secure ML lifecycle, which phase should include validating that the training data has not been tampered with?
Select an answer first
19foundation · easy
Which setting is most directly related to ensuring user data handled by a Gemini Enterprise Agent is not retained longer than necessary?
Select an answer first
20foundation · easy
To protect data processed by a Gemini Enterprise Agent, which security control should be configured?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to PROFESSIONAL-CLOUD-SECURITY-ENGINEER

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Google Cloud. “PROFESSIONAL-CLOUD-SECURITY-ENGINEER” is a trademark of its owner, used for identification only.

3.3 Securing AI Workloads PROFESSIONAL-CLOUD-SECURITY-ENGINEER Practice Questions (Page 4)

Questions 16–20

A security engineer is reviewing threats to a machine learning system. Which attack involves an adversary submitting carefully crafted inputs during the training phase to cause the model to learn incorrect behaviors?

In a secure ML lifecycle, which phase should include validating that the training data has not been tampered with?

Which setting is most directly related to ensuring user data handled by a Gemini Enterprise Agent is not retained longer than necessary?

To protect data processed by a Gemini Enterprise Agent, which security control should be configured?

Keep studying — Ensuring data protection