GCP Generative AI Leader Practice Question 70: Security Command Center AI
Google Cloud Generative AI Leader · Test 7 · question 10 of 10
A CISO is rolling out several gen AI agents across the company and wants a single, organization- wide dashboard that inventories every model, dataset, and AI agent in use, flags agents that have been granted excessive permissions, and detects threats such as prompt injection attempts against those agents, alongside the company's existing cloud security posture.
Which Google Cloud capability should the CISO enable?
Show answer and explanation
Answer:
Security Command Center's AI Protection is built specifically for this: it discovers and inventories AI assets such as models, data sources, endpoints, and agents across an organization, identifies agents that have been granted excessive permissions, and detects AI-specific threats including prompt injection, all alongside the platform's broader cloud security posture management, matching every requirement in the scenario.
- A is wrong because DLP scans data for sensitive patterns like PII; it does not inventory AI agents, flag excessive agent permissions, or detect AI-specific threats like prompt injection. - B is wrong because IAM Recommender optimizes individual human user roles one at a time; it is not an organization-wide AI asset inventory or threat detection capability. - C is wrong because VPC Service Controls creates network-level data exfiltration perimeters; it does not provide an AI asset inventory or agent permission and threat visibility.
Reference: Configure AI Protection