AWS Certified Cloud Practitioner

Guardrails for Amazon Bedrock is the correct feature for setting tone, topic, and safety limits to prevent biased or non-compliant statements.

Explanation:

• Guardrails for Amazon Bedrock provides configurable safety controls that help filter harmful content, block undesirable topics, and enforce compliance requirements

• This feature is specifically designed for financial institutions like banks that need to ensure regulatory compliance and prevent biased outputs

• It allows organizations to define and enforce content policies, including:

  • Blocking specific topics or keywords

  • Setting appropriate tone and language guidelines

  • Filtering out harmful, inappropriate, or non-compliant content

  • Preventing the generation of biased statements

Why other options are incorrect:

• A) Knowledge Bases for Amazon Bedrock: This is for retrieving and using external knowledge sources, not for safety controls

• C) Agents for Bedrock: This enables conversational AI agents to perform tasks, but doesn't provide safety guardrails

• D) Continued pre-training: This refers to additional training of foundation models, not safety controls

For financial institutions dealing with sensitive customer interactions and regulatory requirements, Guardrails for Amazon Bedrock is essential for maintaining compliance and preventing harmful outputs.