AWS Certified Generative AI Developer - Professional AIP-C01

By Amazon · Question Mode

QuestionQ36

AI Safety, Security, and Governance

A financial services company is building a customer-service AI assistant using Amazon Bedrock. The AI assistant must not provide investment advice to users. It must block harmful content, mask personally identifiable information (PII), and retain audit trails for compliance reporting. The AI assistant must apply content filtering to both user inputs and model responses according to content sensitivity.

The company needs an Amazon Bedrock guardrail configuration that effectively enforces these policies with minimal false positives. The solution must support multiple handling strategies for multiple sensitive-content types.

Which solution meets these requirements?

Explanation

Amazon Bedrock Guardrails can combine medium-strength harmful-content filters, denied topics with definitions and example phrases, and sensitive-information filters in one configuration. Sensitive-information policies can apply different actions to inputs and outputs: PII can be anonymized in responses, while financial information can be blocked in inputs. Guardrails evaluate both prompts and model responses, and medium filtering helps limit false positives relative to high-strength filtering.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!