QuestionQ81

VerifiedAI Safety, Security, and Governance

A retail company operates a generative AI (GenAI) product-recommendation application that uses Amazon Bedrock. The application recommends products to customers according to browsing history and demographics. The company must implement fairness evaluation across multiple demographic groups to identify and quantify bias in recommendations between two prompt approaches. The company wants to gather and monitor fairness metrics in real time. The company must receive an alert when the fairness metrics reveal a discrepancy greater than 15% between demographic groups. The company must receive weekly reports that compare the performance of the two prompt approaches.

Which solution meets these requirements with the LEAST custom development effort?

Explanation

Amazon SageMaker Clarify supplies bias metrics that quantify fairness across demographic facets and can monitor bias on an ongoing schedule. Its bias-monitoring workflow produces reports, emits metrics that can be monitored in Amazon CloudWatch, and supports threshold-based alerts. This provides managed analysis and reporting for the demographic disparity requirement while CloudWatch operationalizes monitoring and alerting with less custom logic than post-processing every model response in Lambda.

Learn more

Community Discussion

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