AWS Certified AI Practitioner AIF-C01 Practice Question

Amazon SageMaker Clarify includes a bias analysis feature that produces metrics and visual reports for both pre-training bias (data bias) and post-training bias (prediction bias). This single capability lets teams examine sensitive attributes in the dataset before building a model and then re-evaluate those same attributes on the model's outputs after training. SHAP feature attribution focuses on explainability, Ground Truth handles data labeling, and SageMaker Model Monitor addresses drift after deployment, none of which measure bias both before and after training.