3.3 Securing AI Workloads PROFESSIONAL-CLOUD-SECURITY-ENGINEER Practice Questions (Page 7)

Questions 31–33

1. 31application · medium

   A global law firm is deploying a Gemini Enterprise agent to assist lawyers with drafting contracts. The firm requires that the agent's responses are not used to train the underlying model and that user prompts are not retained beyond the session. What should the administrator configure?

   • AUse a customer-managed encryption key (CMEK) for the Gemini Enterprise agent's data at rest.
   • BDisable data logging for the Gemini Enterprise agent and configure the retention policy to zero days.
   • CSet the data residency to a specific region and enable VPC Service Controls for the Gemini API.
   • DRestrict access to the agent using IAM and require multi-factor authentication for all lawyers.
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   Which activity is part of the monitoring phase of a secure ML lifecycle?

   • AReviewing access logs for the model endpoint
   • BCleaning the training dataset
   • CEncrypting the model artifact
   • DSelecting the ML algorithm
   Reveal Solution Discussion Select an answer first
3. 33foundation · easy

   A data scientist wants to use a dataset containing personally identifiable information (PII) for ML training while reducing the risk of exposing individuals. Which practice directly supports data minimization?

   • AGranting broad access to the dataset
   • BEncrypting the dataset at rest
   • CStoring the dataset in a different region
   • DRemoving unnecessary PII fields from the dataset
   Reveal Solution Discussion Select an answer first