QuestionQ89

VerifiedAI Safety, Security, and Governance

A company uses Amazon Bedrock to create a customer-facing AI assistant that handles sensitive customer inquiries. The company must use defense-in-depth safety controls to block sophisticated prompt-injection attacks, retain audit logs for every safety intervention, and provide cross-Region failover for the AI assistant.

Which solution meets these requirements?

Explanation

Amazon Bedrock Guardrails can block prompt injections by using the PROMPT_ATTACK content filter at HIGH strength. A system-defined guardrail profile enables cross-Region guardrail inference, which routes guardrail evaluations across supported Regions to improve availability. Amazon Bedrock can send model-invocation logs to CloudWatch Logs, and the AWS/Bedrock/Guardrails CloudWatch metrics namespace includes intervention telemetry for guardrail monitoring and auditing. Detect prompt attacks with Amazon Bedrock Guardrails Distribute guardrail inference across AWS Regions Monitor model invocation using CloudWatch Logs and Amazon S3

Learn more

Community Discussion

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