Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
An enterprise is deploying a customer-facing virtual assistant using Gemini on Vertex AI. To adhere to Google's Responsible AI principles and prevent inappropriate or harmful content generation, the engineering team must ensure the model automatically blocks harmful responses such as harassment or dangerous instructions.
Which Vertex AI capability should the team configure to achieve this requirement?
Generative FAQ is an analytical feature within Contact Center Insights that reviews ingested conversation logs to discover trending questions asked by customers and aggregate how those inquiries were handled across past interactions.
While Generative FAQ helps improve general answer accuracy over time, it operates asynchronously on logged conversations rather than acting as a real-time runtime safety filter.
Vertex AI safety settings are built-in governance controls that evaluate prompts and generative responses against specific categories of potential harm. When configuring foundation models like Gemini, developers can assign explicit Harm Categories (such as harassment, hate speech, sexually explicit material, or dangerous content) and define Harm Block Thresholds (such as blocking low-and-above probability harm).
HARM_CATEGORY_HARASSMENT or HARM_CATEGORY_DANGEROUS_CONTENT, the request automatically terminates with a finish reason of SAFETY.Configuring built-in safety settings directly on the generative model provides native, low-latency, and automated enforcement of content safety without requiring external custom filtering pipelines.
Dialogflow utilizes intent classification to match user utterances to predefined conversational intents and triggers webhook fulfillment logic to retrieve data from external systems.
While Dialogflow can orchestrate virtual agents, foundation model safety filtering must be configured directly within the model's safety settings in Vertex AI.
Cloud Identity and Access Management (IAM) manages authentication and authorization across Google Cloud services by specifying who (identity) has what access (roles/permissions) to specific cloud resources.
IAM is an essential pillar of cloud infrastructure security, but content safety governance specifically requires semantic safety filters applied to model inference.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated