Q36Security, Compliance, and Governance for AI Solutions
An e-commerce company is deploying a chatbot. The chatbot will allow users to ask questions about the company’s products and receive details about users’ orders. The company must implement safeguards that filter harmful content from input prompts and chatbot responses. Which AWS feature or resource meets these requirements?
← → navigate · a answer
Community votes
Discussion · 5
A 2
Amazon Bedrock Guardrails are specifically designed to help filter and block harmful, unsafe, or unwanted content in:
. User input prompts
. Model-generated responses
A 2
https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html
A 1
Guardrails protect harmful content
A 1
W/o a doubt, it's Guardrails only. Keyword 'filter harmful contents'..
A 1
Guardrails can:
• Filter toxic, harmful, or unsafe content
• Enforce topic restrictions
• Enforce word/phrase blocking
• Apply to both inputs and outputs
• Work with any Bedrock model
• Provide consistent safety across the application