From the course: Responsible AI on AWS: Bedrock Guardrails, Amazon Q Security, and SageMaker Clarify
Join today to access over 25,000 courses taught by industry experts.
Amazon Bedrock security: Guardrails deep dive
From the course: Responsible AI on AWS: Bedrock Guardrails, Amazon Q Security, and SageMaker Clarify
Amazon Bedrock security: Guardrails deep dive
- [Instructor] Amazon Bedrock implements a multi-layered security architecture that ensures defense in depth through authentication controls, guardrails, and also network security. And each layer operates independently while maintaining secure interfaces between components. Here, we have the authentication layer first that we'll take a look at, and the IAM roles enforce the principle of least privilege. So you hear this all the time with AWS PLP, and in this case it gives you access to the Bedrock API and resources and nothing more. We also have resource policies that are defined from the resource level, for granular access control. We have KMS encryption as well, which provides envelope encryption for data at rest and in transit. And then all authentication methods support long-term credentials and temporary session tokens as well. In the guardrails layer here, we have content filters that implement real-time scanning of prompts and responses. We have input validation that ensures…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.