AI Safety, Security, and Governance II
Explore how to implement AI safety, security, and governance for regulated generative AI applications on AWS. Understand methods for data lineage tracking, audit readiness, transparent decision-making, and compliance with governance standards. Gain insights into AWS services like Bedrock, SageMaker, Glue, and CloudTrail to build secure and auditable GenAI solutions.
We'll cover the following...
Question 43
A financial advisory firm is building a regulated GenAI chatbot using Amazon Bedrock to provide investment insights based on internal research documents. For compliance and audit readiness, the firm must track how data flows from source documents through the RAG pipeline and into foundation model responses. Auditors must be able to trace outputs back to their originating datasets.
Which solution best enables end-to-end data lineage tracking with minimal custom development?
A. Register source datasets in AWS Glue Data Catalog and use metadata tagging to associate documents with downstream RAG components.
B. Store all source documents ...